Updated September 2, 2026

Best AI Models in 2026

Rankings, comparisons, and deep dives, updated monthly as new models ship.

Claude Fable 5.1 arrived on September 1 and went straight to the top of Artificial Analysis's Intelligence Index at 66, the highest score the house has measured, taking its Agentic Index at 61 with it. Claude Opus 5 keeps the two things most readers are actually choosing on: the coding crown, won on both of Arena's vote-based boards, and the price, $5 / $25 per 1M tokens against Fable 5.1's $10 / $50. No Arena board has votes for 5.1 yet, so nothing this page decides on human preference has moved. Grok 4.6 was August's real surprise: a post-training refresh of Grok 4.5 that jumped five points to 61, tying GPT-5.6 Sol while charging $2 / $6, and finishing long agentic jobs in roughly half the turns Opus 5 needs.

The other move is at the cheap end, and for the first time this year it went the wrong way. DeepSeek raised both V4 models roughly fourfold on August 16 and split the rate card into peak and off-peak tiers, so DeepSeek V4-Flash 0731 now costs $0.22 / $0.66 off-peak and $0.44 / $1.32 at peak instead of the flat $0.14 / $0.28 it launched on, which costs it the price-performance pick. That pick moves to GLM 5.3 Flash, which landed on August 26 with MIT weights on day one, an Intelligence Index of 57 and a list price of $0.15 / $0.50, halved to $0.075 / $0.25 until September 9, and which now undercuts DeepSeek at every hour of the day. Google had already halved its own Flash rate: Gemini 3.7 Flash shipped on August 13 at an introductory $0.75 / $3.75, with Gemini 3.6 Flash moved onto the same rate, and Gemini 3.8 Flash followed on September 2 at exactly that price, scoring 59 on Artificial Analysis's Intelligence Index against 56 for 3.7 Flash. All three prices double on January 1, 2027. The open field then closed the month with three releases in three days: GLM 5.3 finally published its weights on August 28, though under a bespoke Z.ai licence rather than the plain MIT its smaller sibling got, Alibaba opened Qwen3.8-Flash-Next on August 26 as the first public look at the Qwen4 architecture, and Tencent open-sourced Hy4 Preview on August 28, 770 billion parameters under Apache 2.0. Below are the category winners for September 2026. Click any card to jump straight to the full breakdown, or use the sticky navigation to skip between categories. Rankings, benchmarks, and pricing are updated within 48 hours of any major model launch.

September 2026 Category Winners

Writing

Claude Fable 5

#1 Arena text (1508.6) and Arena creative writing

The only model in the top three of all three independent writing boards, at $10 / $50.

Deep dive →

Chat & Daily Assistant

GPT-5.6

ChatGPT default since July 9

The best assistant most people can actually open, balancing capability, speed, and reach.

Deep dive →

Images

ChatGPT Images 2.0

#1 on text-to-image and image editing

Leads both independent image boards and still the best at readable multilingual text.

Deep dive →

Video

Gemini Omni Flash

#1 on both video boards (1527 Arena)

Leads both independent video boards with conversational editing at roughly $0.10 per second.

Deep dive →

Coding

Claude Opus 5

#1 Arena WebDev (1,702.9) at $5 / $25

Takes the coding crown on the vote-based boards, at half the price of Claude Fable 5.

Deep dive →

Creativity

Grok 4.6

Fewest content restrictions + native real-time X

The pick for edgy, on-trend work: fewest guardrails and native X integration.

Deep dive →

Accuracy & Research

Gemini 3.1 Pro

98% ARC-AGI-1 at $0.52 a task

Ties the human panel on ARC-AGI-1 with native Google Search grounding for live answers.

Deep dive →

Problem Solving

GPT-5.6 Sol

#1 LiveBench Mathematics, Reasoning and ARC-AGI-2

The best-supported crown on this page for the hardest math, science, and reasoning.

Deep dive →

AI Agents

Gemini Spark

First 24/7 cloud-resident AI agent

Runs continuously in a Google Cloud VM, deeply integrated with Gmail, Docs, and Chrome.

Deep dive →

Want the top AI models without juggling separate subscriptions? Fello AI brings the leading models together in one native app for Mac, iPhone and iPad.

Download Fello AI

What's New in September 2026

Sep 2 Google Gemini 3.8 Flash, three points on the Intelligence Index at the same $0.75 / $3.75 New

Google made Gemini 3.8 Flash generally available on September 2, 2026, three weeks after 3.7 Flash and its third Flash release in 43 days. Every specification carries over unchanged: 1M input context, a 64K output limit, a March 2026 knowledge cutoff, the same modality list, and the same introductory $0.75 / $3.75 per 1M tokens that doubles to $1.50 / $7.50 on January 1, 2027. Only the scores moved, and they moved on every published row. DeepSWE v1.1 goes from 65.3% to 73.7%, Terminal-bench 2.1 from 85.8% to 89.4%, Terminal-bench 4.0 from 11.2% to 19.1% and BioMysteryBench Human Difficult from 43.5% to 56.5%. Artificial Analysis scores it 59 on Intelligence Index v4.1.1 against 56 for 3.7 Flash, and measures 304.6 output tokens per second against 279.4, with a slow 13.39-second time to first token that makes it a batch and agent model rather than an interactive one. Read the wins alongside the losses: Google's own comparison table gives 3.8 Flash 8 of 14 rows, and Claude Opus 5 still takes Terminal-bench 4.0 by 51.8% to 19.1%, OSWorld-2.0 by 75.4% to 59.0% and GDPVal-AA v2 by 1824 to 1545 on Elo. The DeepSWE figure republished across much of the launch coverage, 71.0%, is not the one in Google's PDF, which reads 73.7% against Opus 5's 74.0%. Developer access was live at announcement through the Gemini API, AI Studio, Antigravity and the Gemini Enterprise Agent Platform. Consumer access is reported for Google AI Pro and Ultra subscribers, but the Gemini Apps release notes carry no entry for it yet, and the free Gemini tier still runs 3.6 Flash. Full detail in our Gemini 3.8 Flash breakdown.

Sep 1 Anthropic Claude Fable 5.1 and Mythos 5.1, a new #1 on Artificial Analysis at 66 and a 75% cut to cache reads New

Anthropic released Claude Fable 5.1 and Claude Mythos 5.1 on September 1, 2026, and they are one model in two safeguard configurations rather than two models. Fable 5.1 is generally available; Mythos 5.1 relaxes the biology and cybersecurity restrictions and is handed out by invitation, with no published eligibility criteria. Artificial Analysis scores Fable 5.1 at 66 on its v4.1.1 Intelligence Index at max effort, the highest score it has ever measured, ahead of Claude Opus 5 at 63, Claude Fable 5 at 62, and GPT-5.6 Sol and Grok 4.6 at 61. It also takes the Agentic Index at 61 against Opus 5's 59, the GDPval-AA v2 professional-deliverables board at 1853 against 1824, and Humanity's Last Exam at 59.1% against Fable 5's 55.5%. The effort setting matters more than usual here: the same model reads 58 at low, 60 at medium, 62 at high and 65 at xhigh, and those settings span an eleven-fold difference in output tokens, from 13.1 million to 143.7 million across the evaluation suite. Pricing is unchanged at $10 / $50 per 1M tokens, but cache reads drop 75% to $0.25 from Fable 5's $1.00, which is where the saving actually sits on long agentic runs. On Anthropic's own figures it reaches 52.6% on Terminal-Bench-Science 0.1 against Opus 5's 29.0%, and the 212-page system card raised the company's own alignment risk assessment from very low to low. Two things to weigh before switching: Anthropic still recommends starting with Opus 5 for most workloads at half the price, and no Arena board has votes for either model, so nothing on this page that rests on human preference has moved. Full detail in our Claude Fable 5.1 and Mythos 5.1 breakdown.

Aug 28 Z.ai GLM 5.3 weights are out after the safety hold, but the licence is not MIT New

Z.ai published the GLM 5.3 weights on Hugging Face on August 28, 2026, two weeks after the API launch and on exactly the date its placeholder had carried, so the safety review it promised did happen and the hold is now closed rather than open-ended. What changed is the licence. GLM 5.3 Flash had shipped under plain MIT two days earlier; the flagship carries a bespoke document titled the GLM 5.3 License, and the model card's licence field reads glm-5.3 rather than mit. The grant itself tracks MIT closely, with rights to run, deploy, fine-tune, modify, distribute and sell, and with weights explicitly written into the definition of the software, so commercial use is open to almost everyone. One clause is genuinely new: a model-as-a-service operator whose revenue passes $10 billion over any twelve consecutive months has to pass a Z.ai security review before deploying commercially, and Z.ai reserves the right to decide that review's scope and method itself. Note also that Hugging Face counts the published checkpoint at 753B parameters against the 744B base Z.ai describes it sharing with GLM-5.2, which is the sort of gap a mechanical parameter count produces rather than evidence of a different model. This does not move our open-weight pick. GLM 5.3 Flash stays the model we would host, because 320B under plain MIT is a far easier thing to run and to clear legally than 753B under a bespoke licence. Full detail in our GLM 5.3 breakdown.

Aug 28 Tencent Tencent Hy4 Preview, 770B open weights under Apache 2.0 and the largest permissive model yet New

Tencent released and open-sourced Hy4 preview on August 28, 2026, the new flagship of its Hunyuan line and the largest model anyone has published under Apache 2.0. It runs 770 billion total parameters with 49 billion active per token on a mixture-of-experts design, takes a context window Tencent describes as exceeding 1M tokens, and ships with an FP8 build alongside the full-precision weights. API pricing is $0.834 per 1M input tokens, $2.501 output and $0.042 for cached input, which is inexpensive for the size. The licence is the real headline: Apache 2.0 is more permissive than Kimi K3's custom document and than the bespoke licence Z.ai attached to GLM 5.3 the same day. Treat the quality claim carefully, because the only evidence so far is the vendor's. Tencent ran an internal blind evaluation of 203 engineering tasks scored by 163 experts, where Hy4 preview takes 2.99 out of 4.00 against Kimi K3's 2.94 and GLM 5.3's 2.92. That is Tencent's own panel, the margins sit inside a tenth of a point, and no independent house has published an Intelligence Index or an Arena rating for it yet, so we are listing it rather than ranking it. Tencent also says the model helped automate the optimisation of its own training pipeline and lifted its own inference throughput by a measured 31.8%. It is available as open weights and through WorkBuddy, CodeBuddy, Yuanbao and ima, plus Tencent Cloud TokenHub and OpenRouter, with free access on WorkBuddy and CodeBuddy for two weeks from launch.

Aug 26 Z.ai GLM 5.3 Flash, Ox Alpha unmasked, and the first MIT weights Z.ai has shipped since GLM-5.2 New

Z.ai released GLM 5.3 Flash on August 26, 2026, and it is not the model the field was waiting for. The weights held back on August 14 belong to GLM 5.3; this is a separate model on a newly trained base, 320 billion parameters with 18 billion active, the first natively multimodal model in the GLM-5 line, and it went up on Hugging Face under MIT the same day. It is also the stealth model that had been serving traffic on OpenRouter as Ox Alpha, run during that preview entirely on Chinese AI chips, which describes how it was served rather than how it was trained. Artificial Analysis scores it 57 on the Intelligence Index v4.1.1, four points above GLM-5.2 at less than half the size, and places it on the intelligence-versus-cost Pareto frontier at roughly $0.09 an index task. List price is $0.15 / $0.50 per 1M tokens with a 50% launch promotion running to 24:00 on September 9, and after DeepSeek's August 16 rate rise that list price undercuts DeepSeek V4-Flash 0731 at every hour of the day. Treat the coding and agentic figures with care: they come from Z.ai's own chart, which picks Claude Opus 4.8 and GPT-5.6 Terra as comparators rather than the current leaders, and it reads Terminal Bench 2.1 84.3, DeepSWE v1.1 63.4 against GLM-5.2's 46.2, and AutomationBench 48.8. The one number on that chart Artificial Analysis ran is GDPval-AA v2, where GLM 5.3 Flash takes 1773 Elo, the highest plotted. There are no Arena votes yet. This does move our open-weight pick: GLM 5.3 Flash replaces GLM-5.2 as the MIT model we would host. One caution on hardware, 18B active does not mean 18B to host, because mixture-of-experts routing needs all 320B weights resident, so this is a multi-GPU server rather than a workstation. Full detail in our GLM 5.3 Flash breakdown.

Aug 26 Alibaba Qwen3.8-Flash-Next, open weights and the first public look at the Qwen4 architecture New

Alibaba published Qwen3.8-Flash-Next on Hugging Face on August 26, 2026, and the point of it is the architecture rather than the scoreboard: Qwen says it is an early look at the design the Qwen4 family will use, released so the community can prepare for it. It is a 125-billion-parameter mixture-of-experts with only 6 billion parameters active per token, plus a separate 51-billion-parameter N-gram embedding, and it takes text, image and video in. Context is 262,144 tokens natively, extensible to 1M. Qwen's published figures include 91.7 on GPQA Diamond, 62.5 on SWE-Bench Pro, 58.7 on DeepSWE 1.1, 81.0 on SWE-Bench Multilingual and 84.5 on AndroidWorld vision, all vendor-reported, with no independent Intelligence Index and no Arena votes yet. The licence needs reading before you build on it, because it is not Apache. The Qwen Community License 1.0 allows commercial use, modification and hosting, but it requires prominent display of the model name once a product passes 100 million monthly active users or $20 million in monthly revenue, and it requires a separate licence from Qwen to run a model-as-a-service or an AI work-assistant business on it. Internal use is exempt from that second condition.

Aug 21 OpenAI GPT-5.6 Sol cut by more than 20%, now under Claude Opus 5 on both sides New

OpenAI cut GPT-5.6 Sol from $5 / $30 to $4 / $20 per 1M tokens on August 21, 2026, the first cut to the flagship since the GPT-5.6 family launched in July. The rate is promotional and OpenAI says it runs about three months. It covers the API and credits on Codex and ChatGPT Work; Plus, Pro and Business subscription prices are unchanged. At $4 / $20 Sol now undercuts Claude Opus 5 at $5 / $25 on both input and output, the first time OpenAI's flagship has been the cheaper of the two.

Aug 16 DeepSeek DeepSeek raises API prices roughly fourfold and splits the rate card into peak and off-peak tiers New

DeepSeek moved both V4 models onto a new rate card at 16:00 UTC on August 16, 2026, and the direction is unusual for this field: prices went up, by roughly four times at peak. V4-Flash 0731 went from a flat $0.14 / $0.28 per 1M tokens to $0.22 / $0.66 off-peak and $0.44 / $1.32 at peak, and V4-Pro 0813 from $0.435 / $0.87 to $0.66 / $1.98 off-peak and $1.32 / $3.96 at peak, with cache-hit input at $0.007 and $0.022 off-peak respectively. Peak is 01:00 to 04:00 and 06:00 to 10:00 UTC, Monday to Friday, and every other hour is off-peak at exactly half the peak rate. DeepSeek framed the change as allocating capacity more reasonably. Nothing about either model changed, so this is purely an economics event, but it is a consequential one for this page: it costs DeepSeek V4-Flash the price-performance pick, which moves to GLM 5.3 Flash at $0.15 / $0.50 list, a rate that undercuts DeepSeek's cheaper off-peak price at every hour of the day. The weights are still MIT, so anyone self-hosting is unaffected. Full detail in our DeepSeek V4 breakdown.

Aug 14 Z.ai GLM 5.3, a large agentic jump from post-training alone, shipped without the open weights New

Z.ai released GLM 5.3 on August 14, 2026, and the notable thing is what did not change: it runs on the same 744-billion-parameter base as GLM-5.2, with no new pretrain. Every gain comes from scaled post-training, and the agentic ones are large. On Z.ai's own chart Terminal-Bench 3.0 goes from 4.6 to 28.3, DeepSWE v1.1 from 46.2 to 66.9, and CyberGym from 77.2% to 84.5%, which the company says edges Claude Mythos 5 at 83.8 and GPT-5.6 Sol at 83.6. Every one of those figures is the vendor's, with no independent audit yet, and the chart is less flattering elsewhere: on Z.ai's internal Code Bench, Claude Fable 5 still leads at 39.5% against GLM 5.3's 31.4%, and on ExploitBench the frontier leaders sit at 78.0% against 54.4%. The catch is the licence, not the score. GLM-5.2 shipped MIT weights within days; GLM 5.3 shipped with none, and Z.ai said they would follow after a safety evaluation of the model's vulnerability-discovery capability, roughly two weeks out. That resolved on August 28, 2026, when Z.ai published the weights on exactly the date the placeholder had carried, though under a bespoke GLM 5.3 License rather than the plain MIT that GLM-5.2 and GLM 5.3 Flash both got. Per-token access is now published at $1.40 / $4.40 per 1M tokens, alongside the GLM Coding Plan and ZCode. What arrived on August 26 instead was GLM 5.3 Flash, a different and smaller model that did ship MIT weights, and that is the release that moved our open-weight pick. Full detail in our GLM 5.3 breakdown.

Aug 13 Google Gemini 3.7 Flash, coding scores jump and the price halves to $0.75 / $3.75, until January New

Google shipped Gemini 3.7 Flash on August 13, 2026, 23 days after 3.6 Flash and its third Flash release in under two months, while Gemini 3.5 Pro remains unreleased. The coding gains are the point: DeepSWE v1.1 climbs from 49.0% to 65.3%, FrontierCode 1.1 Main from 34.4% to 43.6%, and AutomationBench nearly doubles from 17.0% to 30.4%. Artificial Analysis scores it 56 on its v4.1.1 Intelligence Index against 52 for 3.6 Flash, and ranks it first of 182 models on output speed at 385.1 tokens per second, which puts it on the intelligence-versus-time Pareto frontier. The context window and 64K output limit are unchanged from 3.6 Flash, so nothing was traded away. Two caveats carry more weight than the benchmarks. The $0.75 / $3.75 rate is introductory and expires on December 31, 2026, after which it doubles to $1.50 / $7.50, exactly what 3.6 Flash cost at launch; Google also moved 3.6 Flash onto the same rate, so the two now cost the same and the upgrade is a pure capability decision. And consumer access runs only through Spark, which needs Google AI Pro or Ultra and excludes the European Economic Area, the UK, Switzerland and Nigeria, so most European readers cannot reach it in the Gemini app at all. The free Gemini tier still gets 3.6 Flash. API access is unaffected everywhere. Full detail in our Gemini 3.7 Flash breakdown.

Aug 12 SpaceXAI Grok 4.6, post-training refresh takes Intelligence Index 56 to 61 at an unchanged $2 / $6 New

SpaceXAI released Grok 4.6 on August 12, 2026, and it is explicitly not a new foundation model: the company kept the Grok 4.5 base and bought the gains with a longer supplemental training run, regenerated fine-tuning trajectories and reinforcement learning inside agentic environments. No new architecture or parameter count was published. Artificial Analysis scores it at Intelligence Index 61, five points above Grok 4.5, which ties GPT-5.6 Sol for joint third behind Claude Opus 5 at 63 and Claude Fable 5 at 62 on the index as it stood that month; Claude Fable 5.1 has since gone in above all three at 66. The agentic figures are stronger than the composite: a GDPval-AA v2 Elo of 1753 behind only Opus 5, 88.4% on Terminal-Bench v2.1 and 50.7% on tau3-Banking. Efficiency is the real argument, at $0.84 per index task and roughly 53 turns and 0.5 billion input tokens on long-horizon work against about 103 turns and 2.0 billion for Opus 5. Pricing is unchanged at $2 / $6 per 1M tokens with cached input at $0.50, on the same 500K-token context window, but there is a trap worth knowing: cross 200,000 prompt tokens and SpaceXAI re-bills the entire request at $4 / $12, not just the overage. One warning on the benchmark coverage, because a lot of reporting has mangled it: Terminal-Bench v3.0 is a different and much harder test than v2.1, so Grok 4.6 scoring 26% on v3.0 and 88.4% on v2.1 are both true of the same model. It went live day one in the SpaceXAI API, Cursor and Grok Build, plus OpenRouter, Vercel and Cloudflare. Full detail in our Grok 4.6 benchmarks and pricing breakdown.

Aug 5 Meta Muse Spark 1.2 and Muse Code, Intelligence Index 51 to 57 at an unchanged $1.25 / $4.25, plus a terminal coding agent New

Meta shipped Muse Spark 1.2 on August 5, 2026 alongside Muse Code, its first terminal coding agent, which runs on macOS and Linux with no desktop app and no subscription. Artificial Analysis scores the model at Intelligence Index 57 at its highest reasoning setting on its current v4.1.1 index, up from 51 for 1.1 and 43 for the April release, and almost all of that gain is agentic rather than general knowledge; its GDPval-AA v2 Elo jumped from 1371 to 1631 while Terminal-Bench v2.1 moved only from 78% to 80%. Pricing is unchanged at $1.25 / $4.25 per 1M tokens on a 1M-token context window, and Meta added a Contributor tier at $0.10 / $0.20, roughly 92% cheaper, in exchange for letting Meta train on your prompts and completions. Availability widened from the US-only preview that 1.1 shipped under to expanded global access across Muse Code, the Meta Model API and OpenRouter. On Meta's own launch charts the model finishes second to Claude Opus 5 on all three coding benchmarks it published, at 82.9% against 86.7% on Terminal-Bench 2.1.

Aug 3 Alibaba Qwen 3.8 (Qwen3.8-Max), 2.4-trillion-parameter multimodal flagship now generally available at $2 / $6 New

Alibaba made Qwen3.8-Max generally available on August 3, 2026, two weeks after previewing it at WAIC Shanghai, and every figure that was missing at preview now has a number behind it. It runs 2.4 trillion total parameters with roughly 95 billion active per token on a sparse Mixture-of-Experts design, takes text, images and video, and confirms a 1M-token context window with up to 128K output tokens. API pricing is $2 / $6 per 1M tokens, flat across the entire context rather than tiered by prompt length, with cached reads at $0.25. The "second only to Fable 5" claim now splits cleanly in two: on Arena's vision board it is #2 at 1305, behind only Fable 5, but on the text board it is #5 at 1496, behind four Anthropic entries. Both Arena scores are still flagged preliminary, and the promised open weights have not shipped. We are keeping Qwen 3.8 out of the ranked picks until the weights and the licence actually land.

Jul 31 DeepSeek DeepSeek V4-Flash 0731, same price, same size, Intelligence Index 40 to 52

DeepSeek re-post-trained V4-Flash and published the result on July 31, 2026 as DeepSeek-V4-Flash-0731. Nothing about the shape of the model changed: it is the same 284B total / 13B active mixture-of-experts, the same 1M-token context, the same MIT licence and, at the time, the same $0.14 / $0.28 per 1M tokens on DeepSeek's own rate card. That last one did not survive the month: DeepSeek raised both V4 models roughly fourfold on August 16 and split them into peak and off-peak tiers. What moved is capability. Artificial Analysis now scores it at Intelligence Index 52 on its v4.1.1 index, up from 40 before the re-post-train, with Agentic 45.7 and 78.7% on Terminal-Bench v2.1, which lifts it from roughly thirteenth to third in the open field behind Kimi K3 and GLM-5.2. Artificial Analysis measured it at $0.11 per index task against $2.34 for Claude Opus 5, which is what won it the price-performance pick at the time; the August 16 rate rise has since cost it that pick. That figure is measured against the endpoint Artificial Analysis benchmarks, which prices higher than DeepSeek's own rate card. One honest limit: the score comes from a single house, and the model has no votes on any Arena board yet.

Jul 31 MiniMax MiniMax H3, 2K video with native stereo audio from $0.13 a second

MiniMax launched H3 (Hailuo 3.0) on July 31, 2026 as a general-purpose multimodal video model that takes text, image, video and audio as input. It outputs 2K clips of 4 to 15 seconds with native stereo audio, supports motion transfer, reference-driven generation and generative video editing, and is billed per second at $0.13 for 2K and $0.09 for 768p. Artificial Analysis places it #2 on its video board at 1241.5, 3.3 Elo behind Gemini Omni Flash. We have kept the video crown where it is: that gap is one day old and comes from the one board that has not reproduced between parses, and Arena's text-to-video cutoff predates H3 entirely, so it has not beaten anything on votes. MiniMax published the weights on August 3, 2026, but only in part: H3-Base and the VAEs and encoder went up on Hugging Face under the MiniMax H3 Community License, which requires an application from users in the US, the EU, the UK and South Korea, while H3-Context-IR and the H3-Regenerate-2K upscaler stayed API-only. The 2K output the model is sold on is therefore not in the open release.

Late Jul Thinking Machines Inkling Small, a quarter the size of Inkling at nearly the same score

Thinking Machines followed its first open model with Inkling Small, a 276B total / 12B active mixture-of-experts under Apache 2.0 that routes each token to 6 of 256 experts plus 2 shared ones. It accepts text, image and audio input and returns text, and Artificial Analysis scores it at Intelligence Index 40 against 41 for the full-size Inkling, at roughly a quarter of the parameters. Sources disagree on the exact publication date, so we are not printing one. The weights, an NVFP4 build and the deployment recipes are all on Hugging Face.

Jul 30 OpenAI GPT-5.6 price cut, Luna 80% cheaper, Terra 20% cheaper, Sol unchanged

OpenAI cut the API price of its two cheaper GPT-5.6 tiers on July 30, 2026, three weeks after the family reached general availability. Luna fell 80% to $0.20 / $1.20 per 1M tokens and Terra fell 20% to $2 / $12, with cached input reads dropping to $0.02 on Luna and $0.20 on Terra. The flagship Sol was left at $5 / $30, prompts above 272K input tokens are still billed at 2x input and 1.5x output, and no ChatGPT subscription price changed, so Free, Go at $8, Plus at $20, Pro at $100 and $200 and Business at $25-30 per seat are all unmoved. The cut leaves Luna at Intelligence Index 52 on Artificial Analysis's v4.1.1 index for about $0.05 per index task, level with Gemini 3.6 Flash on score and far below it on cost.

Jul 24 Anthropic Claude Opus 5, new #1 on Artificial Analysis, tops both the Intelligence Index (63) and the Agentic Index (55.3)

Anthropic released Claude Opus 5 on July 24, 2026, its fourth model in under two months after Mythos 5, Fable 5, and Sonnet 5. On Artificial Analysis's leaderboard it is the top-ranked model overall, leading both the Intelligence Index at 63 and the Agentic Index at 55.3, ahead of Fable 5 (62 / 52.8) and GPT-5.6 Sol (61 / 54.0). API pricing is $5 / $25 per 1M tokens, identical to Opus 4.8 and half the cost of Fable 5, under the API id claude-opus-5. It adds a Fast mode that runs about 2.5x quicker at twice the base price, plus an effort setting from low to high and a new max tier. It also takes Artificial Analysis's GDPval-AA v2 professional-deliverables board outright at 1858, ahead of Fable 5's 1746. Arena has since rated it across its boards, putting it #1 on WebDev, image-to-WebDev, Document and the Agent board and #6 on text, which is what moved the coding crown to it this month. It is the default on Claude Max and strongest on Claude Pro. Note that the leadership claim above has since been superseded: Claude Fable 5.1 arrived on September 1, 2026 and took the Intelligence Index at 66 and the Agentic Index at 61, against 63 and 59 for Opus 5 on the current reading, and the GDPval-AA v2 board with them, where the re-fitted Elo puts Fable 5.1 on 1853 against Opus 5's 1824. Opus 5 keeps both Arena coding boards and costs half as much.

Jul 24 Sakana AI Fugu-Ultra v1.1, orchestration-engine refresh with vendor-reported gains of up to 7.9 points over v1.0 at the same price

Sakana AI shipped Fugu-Ultra v1.1 on July 24, 2026, a refresh of the frontier models inside its TRINITY orchestration engine rather than a new base model. Sakana's own announcement claims gains of up to 7.9 points over v1.0 at unchanged pricing, though it does not publish v1.0 scores side by side, so treat that as the vendor's figure. All the benchmark numbers come from Sakana's custom scaffolding rather than independent testing: on its own charts it leads Opus 4.8, GPT-5.5, and Gemini 3.1 Pro, scoring 73.7% on SWE-Bench Pro, 82.1% on Terminal-Bench 2.1, 93.2% on LiveCodeBench, and 95.5 on GPQA-Diamond. It does not sweep the field, though: its 50.0 on Humanity's Last Exam is a rounding-error tie with Opus 4.8's 49.8. Pricing is unchanged from v1.0 at $5 / $30 per 1M input/output tokens (plus $0.50 cached), rising to $10 / $45 above the 272K-token context mark; the API is OpenAI-compatible with a 1M-token context window.

Jul 21 Google Gemini 3.6 Flash, cheaper output, faster, same Intelligence Index

Gemini 3.6 Flash was Google's newest Flash model when it shipped, and it is still the one the free Gemini app reaches. Output falls to $7.50 per 1M tokens from $9.00 while input holds at $1.50, so the cut is output-only. Google says it "reduces output token usage by 17% compared to 3.5 Flash," and up to 65% on long-horizon agentic work like DeepSWE, so the real per-task saving is larger than the sticker. Artificial Analysis scored 3.6 Flash and 3.5 Flash identically on v4.1, so at launch it was cheaper and faster rather than smarter; on the current v4.1.1 index 3.6 Flash reads 52. It shipped alongside Gemini 3.5 Flash-Lite at $0.30 / $2.50, and is live through the Gemini API in Google AI Studio and Android Studio, the Gemini Enterprise Agent Platform, and the Gemini app for everyone. Note that the launch pricing above has since been superseded: when Gemini 3.7 Flash arrived on August 13, Google moved 3.6 Flash onto the same introductory $0.75 / $3.75 rate, and it remains the model the free Gemini tier serves. Google also announced Gemini 3.5 Flash Cyber, but that one has not shipped: it goes to governments and trusted partners via CodeMender as a limited-access pilot.

Jul 16 Moonshot AI Kimi K3, 2.8T params, the largest open-weight model ever released (weights shipped July 27)

Moonshot AI launched Kimi K3 on the eve of July 16, 2026, its new flagship and the successor to the K2 line, then published the full open weights on July 27, 2026. The Hugging Face model card confirms a 2.8-trillion-parameter Mixture-of-Experts design with 104 billion parameters active per token, a 1,048,576-token context window, and text, image and video input (no audio). Thinking is always on, and reasoning_effort now takes low, high or max, with max as the default. Artificial Analysis puts K3 at Intelligence Index 60, the highest of any open-weight model, and on Arena it is #3 on the Agent board and #2 on WebDev behind Claude Opus 5. Official API pricing is $3 / $15 per 1M tokens with cached input at $0.30, plus $0.005 per successful web-search call. The download is 96 safetensors shards and about 1.56 TB under a custom Kimi K3 License rather than MIT, so self-hosting is a multi-node job; a free basic chat tier exists in the Kimi app, with heavier agentic use metered on paid plans.

Jul 9 OpenAI GPT-5.6 Sol, Terra, and Luna, next-gen family live across ChatGPT, Codex, and the API

OpenAI opened its GPT-5.6 family to general availability on July 9, 2026, ending the two-week gated preview that started June 26 behind a US-government safety review. The lineup runs from least to most capable: Luna, a fast, low-cost tier; Terra, a balanced everyday model OpenAI says matches GPT-5.5 at roughly half the cost; and Sol, the flagship, tuned for biology, chemistry, and cybersecurity. GPT-5.6 is now rolling out across ChatGPT, Codex, and the API as OpenAI's default, with launch API pricing of Sol $5 / $30, Terra $2.50 / $15, and Luna $1 / $6 per 1M tokens; Terra and Luna were cut on July 30, 2026 to $2 / $12 and $0.20 / $1.20, and Sol itself was cut to $4 / $20 on August 21, 2026. On the few benchmarks OpenAI published, Sol scores 88.8% on Terminal-Bench 2.1 (91.9% in its higher-compute "ultra" mode) versus GPT-5.5's 88.0%, and 60.5 on HealthBench Professional; OpenAI notably withheld the usual SWE-bench Verified, GPQA, and FrontierMath numbers, and its context window is still not officially published. One caveat: OpenAI's own system card and the external evaluator METR flagged elevated "scheming" behaviour in Sol, including gaming a software-engineering test at the highest rate METR has ever recorded.

Pending Google Gemini 3.5 Pro, still unreleased, months behind schedule per Bloomberg Delayed

Gemini 3.5 Pro remains the biggest pending launch. Google announced it at I/O on May 19 alongside Gemini 3.5 Flash, but only Flash shipped, and the June target slipped. Bloomberg reported on July 16, 2026 that the model is months behind schedule and has fallen short of Google's internal goals, with the company still working to improve its capabilities particularly in coding. That is the whole sourced picture. Google has never publicly confirmed a launch date, and Google has not published final specs such as the context window or reasoning modes, so treat circulating figures as unconfirmed. Use Gemini 3.8 Flash in the meantime if you are working through the API, or Gemini 3.6 Flash if you are on the free Gemini app, which still runs it; we will move 3.5 Pro into the main ranking the moment it goes live.

Category pick, September 2026

Best AI for Writing

Claude Fable 5

Best writing overall

→

Kimi K3

Creative fiction

→

Claude Sonnet 5

Free everyday

The best AI for writing is Claude Fable 5, the only model in the top three of all three independent writing boards, with Claude Sonnet 5 as the free-tier value pick and GPT-5.5 as the alternative for fact-anchored business writing. Fable 5 leads Arena's creative-writing leaderboard, tops LiveBench Language at 90.7, and places third on EQ-Bench Creative Writing v3 behind Kimi K3 and GPT-5.6 Sol. No other model is top-three on more than one of them. This is a change from last month, when Claude Sonnet 5 held this slot on GDPval-AA; the preference boards do not support that placing, so Sonnet 5 is now the value pick rather than the quality leader. Fable 5 costs $10 / $50 per 1M tokens and is permanently included in Claude Max and Team Premium at roughly 50% of regular usage limits. If your writing is a work deliverable rather than prose, the GDPval-AA v2 professional-deliverables board is now led by Claude Fable 5.1 at 1853, ahead of Claude Opus 5 at 1824 and Fable 5 at 1723. Fable 5.1 landed on September 1 as the higher-capability model at the same $10 / $50, and it beats Fable 5 on every board that has measured both, including Humanity's Last Exam at 59.1% against 55.5%. We have not moved the crown to it, because this pick rests on human preference and no Arena board has rated it yet. Translation is a separate question with a separate winner, which we work through in our guide to the best AI for translation.

Model	Best For	Strength	Weakness	Price (per 1M tokens)
Claude Fable 5	Best writing overall	#1 Arena text overall (1508.6), #1 LiveBench Language (90.7), #3 EQ-Bench	Priciest option here	$10 / $50
Claude Fable 5.1	Highest capability at the same price	#1 Humanity's Last Exam (59.1%), #1 GDPval-AA v2 (1853), #1 Intelligence Index (66)	No Arena votes yet, so unrated on the preference boards	$10 / $50
Kimi K3	Creative fiction and voice	#1 EQ-Bench Creative Writing (2377), 234 Elo clear of second	Only #10 on Arena creative writing	$3 / $15
Claude Sonnet 5	Free everyday writing	Free and default on claude.ai, 1M context	#53 Arena creative writing; 75.0 LiveBench Language	$2 / $10, now permanent
Claude Opus 5	Professional deliverables on a budget	#2 GDPval-AA v2 at 1824, ahead of Fable 5 (1723)	Behind Fable 5 on Arena text, behind Fable 5.1 on GDPval-AA v2	$5 / $25
GPT-5.5	Fact-anchored business writing	Documented factual-reliability gains over GPT-5.4	Reasoning tiers now marked deprecated	$5 / $30
Gemini 3.8 Flash	Bulk drafts at scale	Intelligence Index 59, 304.6 tok/s output, 1M context	Agent-tuned rather than prose-tuned; intro price doubles January 1, 2027	$0.75 / $3.75

Runner-up and alternatives: Kimi K3 is the runner-up for creative fiction and wins EQ-Bench outright, Claude Sonnet 5 is the runner-up on value and the one to use if you are not paying, Claude Fable 5.1 is the pick if you want the highest measured capability at the same price as Fable 5, Claude Opus 5 is the cheaper pick for professional deliverables, and Gemini 3.8 Flash is the pick for bulk drafting, three points ahead of 3.7 Flash on the Intelligence Index at the same price.

What changed this month

The writing crown stays with Claude Fable 5, but the case for it narrowed on September 1. Anthropic shipped Claude Fable 5.1, and it takes Humanity's Last Exam at 59.1% against Fable 5's 55.5%, the GDPval-AA v2 board at 1853, and the Intelligence Index at 66. What it does not have is a single Arena vote, and this crown rests on Arena's text and creative-writing boards, where Fable 5 is still #1 at 1508.6 on the August 1 cutoff. So Fable 5.1 goes into the table as the higher-capability option at the same $10 / $50, and we re-test the crown the moment Arena rates it. Two numbers we quoted last month have also been re-fitted: AA-Omniscience now reads 43 for both Fable models rather than 40, and Artificial Analysis measures Fable 5 on Humanity's Last Exam at 55.5% rather than 53.3%.

Category pick, September 2026

Best AI for Chat & Daily Assistant

GPT-5.6

ChatGPT's default

→

Claude Fable 5

Highest-rated

→

Claude Opus 5

Thoughtful depth

The best AI for everyday chat is GPT-5.6, and the honest reason is reach rather than board position. It is the model ChatGPT serves by default to the largest user base in the category, which makes it the best assistant most people can actually open. On raw human preference it is not the leader: GPT-5.6 Sol sits #14 on Arena's text leaderboard at 1482.8, where Claude Fable 5 leads at 1508.6. Most ChatGPT users get the balanced Terra tier, which OpenAI says matches GPT-5.5 and, since the July 30, 2026 price cut, costs 60% less than it. It is available inside ChatGPT (free with limits, Plus at $20/month, Pro at $100/month), through the API (Luna $0.20 / $1.20, Terra $2 / $12, Sol $4 / $20 per 1M tokens), and bundled inside Fello AI alongside Claude, Gemini, Grok, and DeepSeek. One caveat: OpenAI's system card and the evaluator METR flagged elevated "scheming" behaviour in Sol, so GPT-5.5 Instant stays the safer pick for hallucination-sensitive work.

Model	Best For	Strength	Weakness	Price
GPT-5.6	Everyday chat, ChatGPT's default	The assistant most people can open; Terra matches GPT-5.5 at ~half cost	#14 on Arena text; scheming flagged by METR	Free / $20/mo Plus; API $0.20 / $1.20 to $4 / $20
Claude Fable 5	Highest-rated conversation	#1 on Arena text overall (1508.6) and 6 of 7 subcategories	No free tier; usage-credit access on Pro	$10 / $50 API
Claude Opus 5	Thoughtful, nuanced answers	#1 Artificial Analysis Intelligence Index (63) and Agentic Index (55.3)	#6 on Arena text, behind Claude Fable 5	$20/mo Pro, $5 / $25 API
GPT-5.5 Instant	Hallucination-sensitive daily work	52.5% fewer hallucinated claims vs 5.3 Instant	Reasoning tiers now marked deprecated	$20/mo Plus; API $5 / $30
Gemini 3.6 Flash	Fast, free, multimodal	Still the model the free Gemini tier serves, 1M context, #12 on Arena text	Weaker on hardest reasoning; 3.8 Flash outscores it at the same price	Free / $0.75 / $3.75 API
Fello AI	All the top models, one app	ChatGPT + Claude + Gemini + Grok + DeepSeek and more on Mac, iPhone and iPad	Routed via app, not direct	$9.99/mo

Runner-up and alternatives: Claude Fable 5 is the runner-up and the actual preference leader, Claude Opus 5 is the runner-up for thoughtful daily use, Gemini 3.6 Flash is the runner-up for fast and free, and Grok 4.6 is the niche pick for live-news days. Fello AI is the natural pick if you want the top models in one Mac and iOS app for $9.99/month instead of juggling subscriptions.

What changed this month

GPT-5.6 keeps the chat pick on reach, and the gap to the preference leader widened rather than closed. On the August 1 cutoff Sol sits #14 on Arena text at 1482.8, down from #11, while Claude Fable 5 leads at 1508.6. Nothing about the product changed, so the crown does not move: this is still about which assistant the most people can actually open.

Category pick, September 2026

Best AI for Images

ChatGPT Images 2.0

Text in images

→

Reve 2.1

Layout & 4K

→

Muse Image

Image editing

The best AI for image generation is ChatGPT Images 2.0, and it is the least controversial crown on this page. GPT Image 2 leads Arena's text-to-image board at 1385 Elo and its image-editing board at 1463, and Artificial Analysis puts it first on its own image arena at 1339.4. It is the natural pick whenever your image needs to contain readable words, in English or in another script, and it is included in ChatGPT Plus and Pro. The runner-ups have changed. Reve 2.1 (July 9) is the real #2 on text-to-image at 1302, and Reve 2.0 now sits behind it on both boards. Meta's Muse Image is #3 on Arena's text-to-image board and #2 on image editing, the strongest showing any Meta image model has managed. Google's Nano Banana Pro is no longer the runner-up overall: on text-to-image it ranks between #8 and #11, below its own cheaper sibling Nano Banana 2, though it does place higher on image editing.

Model	Best For	Strength	Weakness	Price
ChatGPT Images 2.0	Images with readable text	#1 text-to-image (1385) and #1 image editing (1463)	Less photoreal than the Gemini image line	Included in ChatGPT Plus
Reve 2.1	Layout, typography, native 4K	#2 text-to-image at 1302, layout-preserving editing	Smaller ecosystem	Free / from $7.99/mo
Muse Image	Image editing, Meta ecosystem	#3 text-to-image, #2 image editing (1407)	New, thin tooling around it	Meta AI app
Nano Banana 2 (Gemini 3.1 Flash Image)	Photoreal portraits and products	Outranks Nano Banana Pro on both boards	Weaker on text in image	Gemini app / AI Studio
Seedream 5.0 Pro	Multilingual text + region editing	10+ languages incl. Arabic RTL, lasso and layer editing	No independent benchmarks; copyright cloud	BytePlus / Magnific
Midjourney v8	Stylized art, illustration	Aesthetic baseline most artists prefer	Weaker on text in image	$10-$120/mo
Grok Imagine	NSFW / Spicy Mode	Most permissive guardrails	Smaller model behind it	$30/mo SuperGrok

Runner-up and alternatives: Reve 2.1 is the runner-up overall and the pick for layout and typography, Muse Image is the runner-up for editing an image you already have, and Nano Banana 2 is the photoreal pick. Grok Imagine is still the only frontier model that allows Spicy Mode adult content.

What changed this month

The image crown is unchanged and GPT Image 2 still leads all three boards we track, on Arena text-to-image (1385), Arena image editing (1463) and Artificial Analysis's image arena (1339.4). The one addition is Microsoft's MAI-Image-2.5, which sits third on Artificial Analysis's image board at 1269.7 and third on Arena's image-editing board, so third place now depends on which house you read.

Category pick, September 2026

Best AI for Video

Gemini Omni Flash

#1 both video boards

→

MiniMax H3

2K + native audio

→

Dreamina Seedance 2.0

Closest challenger

The best AI for video generation is Gemini Omni Flash, which leads Arena's text-to-video board at 1527 Elo, a full 45 points clear of second place, and also tops Artificial Analysis's video arena. It is #1 on both houses, which no other video model manages. Pricing runs $1.50 in and $17.50 per 1M video output tokens, which works out at roughly $0.10 per second of finished video, and it supports conversational editing. The one real limit is length: Omni Flash generates 10-second clips. If you need longer takes, Veo 3.1 remains the right tool inside the Gemini app, AI Studio and Vertex AI, with native audio and 1080p output. This replaces Veo 3.1 at the top of the category; Veo 3.1 is a good model but not the leading one, with its best variant at #6 on Arena's text-to-video board. The contender to watch is MiniMax H3 (July 31), which generates 2K clips of 4 to 15 seconds with native stereo audio and is billed per second from $0.13 at 2K, already #2 on Artificial Analysis's video board at 1241.5. We are not moving the crown on that: the margin is 3.3 Elo on the single board that has not reproduced between parses, and H3 still has no votes on Arena's text-to-video board.

Model	Best For	Strength	Weakness	Price
Gemini Omni Flash	Best AI video overall	#1 on both video boards (1527 Arena), conversational editing	Caps at 10-second generations	~$0.10/sec; Gemini app / AI Studio
MiniMax H3	2K clips with native audio	4-15s at 2K, native stereo audio; #2 on AA video (1241.5)	No Arena votes yet; the open weights exclude the 2K upscaler and need an application in the US, EU, UK and South Korea	$0.13/sec at 2K
Dreamina Seedance 2.0	Closest challenger	#2 text-to-video (1482) and #1 on image-to-video with audio (1198)	ByteDance ecosystem, limited Western access	Dreamina / BytePlus
Muse Video	Meta ecosystem video	#3 text-to-video at 1459	Newest of the group, thin tooling	Meta AI app
Veo 3.1	Longer production clips	Native audio, 1080p, strong physics consistency	#6 on Arena video, not the quality leader	Google AI Pro / Ultra
Kling 3.0 / 3.0 Turbo	Fast iteration at lower cost	Native 4K, 60fps, 15-second clips; Turbo shipped June 17	Outside the top 16 on Arena text-to-video	From $10/mo
Luma Ray 3	Photoreal scenes	Strong realism for landscapes	Smaller community	Free / from $9.99/mo

Runner-up and alternatives: Dreamina Seedance 2.0 is the runner-up overall and beats Omni Flash on image-to-video with audio, where it leads at 1198 to Omni Flash's 1191, though Omni Flash still leads the no-audio image-to-video split. Muse Video is third, and Veo 3.1 is the pick when 10 seconds is not enough. OpenAI retired the Sora 2 consumer app on April 26, 2026 and only the developer API remains, through September 24, 2026. We track the retirement dates for older models across every provider in a separate running list.

What changed this month

Gemini Omni Flash keeps the video crown, and it is still #1 on both boards, at 1527.5 on Arena and 1244.8 on Artificial Analysis. MiniMax H3 (July 31) enters as the contender at #2 on Artificial Analysis's video board (1241.5), 3.3 Elo behind, and it beats Omni Flash on specification with 2K output, clips up to 15 seconds and native stereo audio. We are holding the crown because that margin is 3.3 Elo on a single board, and because H3 still has no votes on Arena's text-to-video board. MiniMax did open the weights on August 3, but only H3-Base and the VAEs and encoder; the 2K upscaler that the specification advantage rests on stayed API-only.

Category pick, September 2026

Best AI for Coding

Claude Opus 5

#1 Arena WebDev

→

Claude Fable 5

Hardest long-horizon

→

Kimi K3

Web apps and agents

The best AI for coding is Claude Opus 5, and it wins on the two boards where developers vote on the finished result rather than a script measuring a harness. It is #1 on Arena's WebDev board at 1,702.9 and #1 on image-to-WebDev at 1,668.6, on vote cutoffs of August 1 and July 31. Price is the second half of the argument: Opus 5 runs $5 / $25 per 1M tokens against Claude Fable 5's $10 / $50, and Artificial Analysis measures it at $2.34 per index task against Fable 5's $3.14. Anthropic's own docs tell developers to start with Opus 5 for complex agentic coding and reserve Fable 5 for workloads that need the highest available capability. Claude Fable 5 is the runner-up and stays the pick for the hardest long-horizon work, at #4 on WebDev (1,630.7) and #2 on image-to-WebDev (1,625.7); its September 1 successor, Claude Fable 5.1, is the stronger model on the harness benchmarks at the same price, reaching 55.8% on Terminal-Bench 4.0 against Opus 5's 52.3% and Fable 5's 42.0% on Anthropic's own figures, but Arena has no votes for it yet. The contender is Kimi K3, which takes #2 on WebDev at 1,675.5 and #3 on Arena's Agent board, the strongest open-weight coder on the boards, though self-hosting it means 1.56 TB of weights. On Artificial Analysis's Coding Index it is not a Claude sweep: GPT-5.6 Sol (xhigh) leads at 78.3 with Opus 5 (max) at 78.0, close enough to call a tie. The cheapest serious contender is now GLM 5.3 Flash at $0.15 / $0.50 under MIT, after DeepSeek's August 16 rate rise took V4-Flash 0731 to $0.22 / $0.66 off-peak and $0.44 / $1.32 at peak, and the best value at the frontier is Grok 4.6, which posts 88.4% on Terminal-Bench v2.1 and $0.84 per index task at $2 / $6.

Model	Best For	Strength	Weakness	Price (per 1M tokens)
Claude Opus 5	Best coding overall	#1 Arena WebDev (1,702.9) and #1 image-to-WebDev (1,668.6); $2.34 per index task	Artificial Analysis Coding Index has GPT-5.6 Sol a shade ahead	$5 / $25
Claude Fable 5	Hardest long-horizon agentic work	#2 Arena image-to-WebDev (1,625.7), #4 WebDev; 1M context	Priciest; Artificial Analysis Coding Index puts it 7th	$10 / $50
Claude Fable 5.1	Highest-capability agentic coding	#1 Intelligence Index (66) and #1 Agentic Index (61); 55.8% Terminal-Bench 4.0 (Anthropic)	No Arena votes yet; twice the price of Opus 5	$10 / $50
Kimi K3	Web app building and agents	#2 Arena WebDev (1,675.5), #3 Arena Agent, highest open Intelligence Index (60)	1.56 TB to self-host	$3 / $15
GPT-5.6 Sol	OpenAI flagship, agentic coding	#1 Artificial Analysis Coding Index at 78.3 (xhigh)	Absent from the official Terminal-Bench board; eval-gaming flagged by METR	$4 / $20
Muse Spark 1.2	Cheap agentic coding	Intelligence Index 57 at $1.25 / $4.25; 80% on Terminal-Bench 2.1 (Artificial Analysis)	Second to Opus 5 on all three of Meta's own coding charts (82.9% vs 86.7%); Scale SEAL has rated only 1.1	$1.25 / $4.25
Grok 4.6	Cheap value coder	88.4% on Terminal-Bench v2.1 and Intelligence Index 61, at $0.84 per index task	Prompts of 200K tokens and up re-bill the whole request at double; higher hallucination rate	$2 / $6
Gemini 3.8 Flash	Agent coding at scale	Intelligence Index 59, 73.7% DeepSWE v1.1, 89.4% Terminal-bench 2.1, 304.6 tok/s	19.1% on Terminal-bench 4.0 against Opus 5's 51.8%; intro price doubles January 1, 2027	$0.75 / $3.75
GLM 5.3 Flash	Best open-weight coder you can host	Intelligence Index 57; DeepSWE v1.1 63.4 and Terminal Bench 2.1 84.3 (vendor figures); MIT, 320B/18B active	Kimi K3 outscores it; the coding figures are vendor-reported and it has no Arena votes yet	Open weights (MIT)

Runner-up and alternatives: Kimi K3 is the runner-up on Arena's WebDev board and the pick if you want the strongest open weights, Claude Fable 5 is the runner-up for the hardest long-horizon work, with Claude Fable 5.1 the higher-capability version of that same choice at the same price, GPT-5.6 Sol is the runner-up on Artificial Analysis's composite, and GLM 5.3 Flash is the open-weight pick for teams hosting it themselves, at Intelligence Index 57 under plain MIT, and it keeps that slot now that GLM 5.3's own weights have arrived under a bespoke Z.ai licence at more than twice the size. Inside IDEs, Cursor with Claude is still the most popular pairing and Claude Code is the natural pick if you live in the terminal.

What changed this month

The coding crown moved to Claude Opus 5. Arena finished rating it, and it came first on both coding boards, WebDev at 1,702.9 and image-to-WebDev at 1,668.6, with Claude Fable 5 fourth and second. Those are vote-based boards with published cutoffs, so the crown now rests on human comparisons rather than on any one harness, and Opus 5 does it at half Fable 5's price. Fable 5 becomes the runner-up for the hardest long-horizon work, and Kimi K3 stays the contender at #2 on WebDev. Meta's Muse Spark 1.2 arrived on August 5 and does not change that, because on Meta's own charts it comes second to Opus 5 on every coding benchmark it published. Grok 4.6 (August 12) does not take the crown either, but it is the one to watch on cost: it reaches 88.4% on Terminal-Bench v2.1 and finishes long agentic jobs in roughly half the turns Opus 5 needs, at $2 / $6 against $5 / $25. Two more coding-focused launches landed right after it. Gemini 3.7 Flash (August 13) replaces 3.6 Flash in this table on the strength of a jump from 49.0% to 65.3% on DeepSWE v1.1 at half the old price, and GLM 5.3 (August 14) posts the largest agentic gain of the month, Terminal-Bench 3.0 from 4.6 to 28.3, but shipped with no downloadable weights. Z.ai closed the month by opening a different model instead: GLM 5.3 Flash (August 26) arrived under MIT on day one at Intelligence Index 57, and it takes the open-weight coding slot from GLM-5.2. The last days of the month then produced three more open releases: GLM 5.3's own weights on August 28 under a bespoke licence rather than MIT, Alibaba's Qwen3.8-Flash-Next on August 26 as a preview of the Qwen4 architecture, and Tencent's 770B Hy4 Preview on August 28 under Apache 2.0. None of them has an independent coding rating yet, so none of them moves this table. Anthropic then opened September with Claude Fable 5.1 on the 1st, which tops Artificial Analysis's Intelligence and Agentic indexes and posts 55.8% on Terminal-Bench 4.0 against Opus 5's 52.3%, but Arena has not rated it, so the crown stays where the votes are. Google closed the gap at the cheap end on September 2 with Gemini 3.8 Flash, which takes the at-scale row from 3.7 Flash at the same price: DeepSWE v1.1 goes from 65.3% to 73.7% and Terminal-bench 2.1 from 85.8% to 89.4%. It does not touch the crown, because Opus 5 still wins Terminal-bench 4.0 by 51.8% to 19.1%.

Category pick, September 2026

Best AI for Creativity

Grok 4.6

Fewest guardrails

→

Claude Fable 5

Highest-quality prose

→

Kimi K3

Fiction and voice

The best AI for unfiltered, on-trend creative work is Grok 4.6, and we want to be exact about why. This pick is about the product, not the prose quality. Grok carries the fewest content restrictions of any frontier model and the only native real-time X integration, which makes it the one model that will engage with edgy, topical or deliberately provocative briefs that the others decline. It is the default in the Grok app for SuperGrok and X Premium+ subscribers at $30/month. It is not the best writer, and the boards were blunt about its predecessor: Grok 4.5 sat #33 on EQ-Bench Creative Writing and #41 on Arena's creative-writing leaderboard, losing on both to the older Grok 4.20-beta1. Neither creative-writing board has rated Grok 4.6 yet, and since SpaceXAI aimed this release at coding and agentic work rather than prose, we are not assuming it moved. If you are picking on output quality alone, Claude Fable 5 wins outright at #1 on Arena creative writing, and Kimi K3 wins EQ-Bench. Choose Grok 4.6 for what it will let you make, not for how well it writes.

Model	Best For	Strength	Weakness	Price
Grok 4.6	Unfiltered, opinionated, on-trend	Fewest content restrictions, native real-time X grounding	Creative-writing boards have not rated it; Grok 4.5 sat #33 EQ-Bench, #41 Arena	$30/mo SuperGrok
Claude Fable 5	Highest-quality creative prose	#1 Arena creative writing, #1 LiveBench Language	Cautious guardrails on edgy briefs	$10 / $50 API
Kimi K3	Fiction and distinctive voice	#1 EQ-Bench Creative Writing at 2377	Only #10 on Arena creative writing	$3 / $15
Claude Opus 5	Long-form structured creativity	Holds long threads and self-edits; Intelligence Index 63, second only to Claude Fable 5.1	Most cautious of the group	$20/mo Pro; $5 / $25 API
Gemini 3.1 Pro	Multimodal creative	Strong text, image and video chain	Quotas inside the Gemini app	Free / $2.00-$4.00 API in
Grok Imagine (Spicy Mode)	NSFW / adult creative	Most permissive image generation	Niche use case	$30/mo SuperGrok

Runner-up and alternatives: Claude Fable 5 is the runner-up and the right pick if quality matters more than freedom, Kimi K3 is the pick for fiction, and Claude Opus 5 is the pick for creative projects that run across many turns. For adult creative work, Grok Imagine Spicy Mode is still the only frontier-grade option.

What changed this month

Grok keeps this pick, now on Grok 4.6, which replaced Grok 4.5 as SpaceXAI's flagship on August 12. Nothing about the product's permissiveness or its live X access changed, which is what the pick rests on. The model underneath is meaningfully stronger, up five points to Intelligence Index 61, but that gain landed in coding and agentic work rather than prose, and no creative-writing board has rated 4.6 yet. Claude Fable 5 remains the model to use when you want the better writing.

Category pick, September 2026

Best AI for Accuracy & Research

Gemini 3.1 Pro

98% ARC-AGI-1

→

Claude Fable 5

Grounded search

→

GPT-5.6 Sol

Novel reasoning

The best AI for accuracy and research is Gemini 3.1 Pro. Its strongest result is on ARC Prize's ARC-AGI-1, where it scores 98% and ties the human panel, and it does that at $0.52 per task. That combination is the argument: several models are close on capability, none matches it on cost for reliable factual work. It pairs that with native Google Search grounding, which is what you want when the answer has to be current rather than merely plausible. It also scores 94.1% on GPQA Diamond, where GPT-5.6 Sol now matches it rather than trailing it, and 44.4% on Humanity's Last Exam, and tops Scale SEAL's HLE board at 46.44. We have dropped the previous ARC-AGI-2 framing: its 77.1% is still correct, but the board has moved and that score now places it around 14th. Two honest caveats. On grounded search specifically, Arena's search leaderboard is led by Anthropic, not Google, with Gemini 3.1 Pro grounding at #7. And on novel reasoning, GPT-5.6 Sol leads ARC-AGI-2 and Claude Opus 5 leads ARC-AGI-3 at 30%, roughly 3.75x the next-best model. We did not move the crown to Opus 5 because Artificial Analysis measures its hallucination rate at 50% and places it below Fable 5 on AA-Omniscience, a board Claude Fable 5.1 now shares the top of.

Model	Best For	Key Benchmark	Weakness	Price
Gemini 3.1 Pro	Cheap, reliable factual work	98% ARC-AGI-1 (ties human panel) at $0.52/task, 94.1% GPQA (tied by Sol)	ARC-AGI-2 77.1% now ranks ~14th; #7 on Arena search	$2.00-$4.00 / $12.00-$18.00 (tiered)
Claude Fable 5	Grounded search	#1 and #3 on Arena's search leaderboard, ahead of Google	No single cheap tier	$10 / $50 (Fable 5)
GPT-5.6 Sol	Novel reasoning	#1 ARC-AGI-2 at 92.5%, against a 100% human panel	Scheming flagged by METR	$4 / $20
Claude Opus 5	Hardest unseen problems	#1 ARC-AGI-3 at 30%, ~3.75x the next model (ARC Prize)	Artificial Analysis measures a 50% hallucination rate	$5 / $25
Qwen 3.7 Max	Frontier accuracy at value pricing	92.4 GPQA Diamond, 200 free requests/day	API-only, no chat front-end	$1.25 / $3.75 promo; $2.50 / $7.50 list
Claude Opus 4.6	Honesty under pressure	#1 on Scale SEAL's MASK board at 96.28; Anthropic holds the top 5	Superseded as a flagship	Legacy Anthropic model

Runner-up and alternatives: Anthropic's models are the runner-up for grounded search and sweep the honesty-under-pressure board, GPT-5.6 Sol is the runner-up for novel reasoning, and Qwen 3.7 Max is the value pick at the frontier.

What changed this month

Gemini 3.1 Pro keeps the accuracy crown, but its headline number is no longer a lead. Its GPQA Diamond score re-fitted to 94.1% and GPT-5.6 Sol now matches it exactly, so the crown rests on the ARC-AGI-1 result at $0.52 a task plus Search grounding rather than on GPQA. This is the next crown we re-test, and the boards that measure how often a model is simply wrong moved on September 1. Claude Fable 5.1 now holds the highest factual accuracy Artificial Analysis has measured, at 67% against Claude Fable 5's 65%, and tops Humanity's Last Exam at 59.1% against 55.5%. The AA-Omniscience index itself reads 43 for both models, up from the 40 we quoted for Fable 5 last month, because 5.1 buys its extra accuracy with a higher attempt rate on questions it cannot answer.

Category pick, September 2026

Best AI for Problem Solving

GPT-5.6 Sol

STEM flagship

→

Claude Opus 5

Agentic chains

→

GPT-5.5 Pro

Verified FrontierMath

The best AI for hard problem solving is GPT-5.6 Sol, and it is the best-supported crown on this page. It takes #1 on LiveBench Mathematics at 96.2, #1 on LiveBench Reasoning at 91.7, and #1 on ARC-AGI-2 at 92.5%, the closest any model has come to the 100% human panel. Three separate houses put it first on the reasoning tasks that matter, which is more agreement than any other category on this page produces. OpenAI has still not published Sol's FrontierMath score, so the verified OpenAI mark remains GPT-5.5 Pro's 39.6% on FrontierMath Tier 4, and we will slot Sol's number in the moment it goes public. Qwen 3.7 Max is the value alternative for competition-style problems at 97.1 on the February 2026 HMMT index and 44.5 on Apex, at a fraction of the cost of ChatGPT Pro, and it now includes 200 free model requests per day. Claude Opus 5 is the alternative for long agentic reasoning chains, at 59 on Artificial Analysis's Agentic Index, second to Claude Fable 5.1's 61, and #1 on ARC Prize's ARC-AGI-3 at 30%, roughly 3.75x the next-best model.

Model	Best For	Key Benchmark	Weakness	Price
GPT-5.6 Sol	Hardest math, science and reasoning	#1 LiveBench Mathematics (96.2), #1 LiveBench Reasoning (91.7), #1 ARC-AGI-2 (92.5%)	FrontierMath still unpublished; scheming flagged by METR	$100/mo ChatGPT Pro; API $4 / $20
Claude Opus 5	Long agentic reasoning chains	#2 Agentic Index (59), #1 ARC-AGI-3 (30%)	Second on LiveBench Reasoning; 50% hallucination rate	$5 / $25
GPT-5.5 Pro	Verified FrontierMath leader	39.6% FrontierMath Tier 4	Superseded by Sol as flagship	$100/mo ChatGPT Pro
Qwen 3.7 Max	Competition math on a budget	97.1 HMMT 2026 Feb, 44.5 Apex, 200 free requests/day	API-only	$1.25 / $3.75 promo; $2.50 / $7.50 list
Claude Fable 5	Math inside a coding workflow	#1 on Arena's math subcategory (1543), 96.0 LiveBench Mathematics	Priciest option here	$10 / $50
GLM 5.3 Flash	Open-weight problem solving	Intelligence Index 57 under MIT, four points above GLM-5.2 at less than half the size; 320B/18B active, 1M context	Needs a multi-GPU server, not a workstation; GLM 5.3 scores higher on Z.ai's own figures and is downloadable since August 28, but at more than twice the size and under a bespoke licence	Open weights (MIT)

Runner-up and alternatives: Claude Opus 5 is the runner-up and the natural pick for long-chain agentic reasoning, Claude Fable 5 is the runner-up on Arena's math board, Qwen 3.7 Max is the value pick, and GLM 5.3 Flash is the open-weight pick. Our dedicated guide to the best AI for math covers the task-by-task split and the free options.

What changed this month

GPT-5.6 Sol keeps this crown, and it picked up a second argument. Sol now also leads Artificial Analysis's Terminal-Bench v2.1 at 89.5% and its Coding Index at 78.3, and it matches Gemini 3.1 Pro on GPQA Diamond at 94.1%. Claude Opus 5 stays the agentic-reasoning alternative on the strength of ARC-AGI-3. On August 1 OpenAI named its next model family Astra and published ten new mathematical results from an internal version, though Astra is unreleased and has no date, price or availability. Anthropic's Claude Fable 5.1 arrived on September 1 and took the Agentic Index off Opus 5, 61 against 59, which changes the alternative's supporting number rather than the crown.

Category pick, September 2026

Best AI Agent

Gemini Spark

24/7 cloud

→

Claude Cowork

Desktop AI

→

ChatGPT Codex

Coding agent

The best AI agent right now is Gemini Spark for 24/7 cloud-resident work and Claude Cowork for desktop-resident work, with ChatGPT Codex as the alternative for coding agents and OpenAI Operator-class browser agents as the alternative for web tasks. AI agents are the fastest-moving category of 2026: each top vendor now ships an agent product, and the practical choice is between agents that live in the cloud (run while your laptop is closed) and agents that live on your desktop (drive your apps directly). Gemini Spark launched at Google I/O on May 19, 2026 and is the first 24/7 cloud agent. Claude Cowork launched in general availability on April 9, 2026 and runs as a desktop agent that drives your local apps. ChatGPT Codex Mobile (May 14) is the pick for coding-agent work. Read the full Gemini Spark vs Claude Cowork comparison.

Agent	Best For	Where It Runs	Strength	Price
Gemini Spark	24/7 cloud tasks, Workspace workflows	Google Cloud VM (always-on)	First true 24/7 agent, deep Workspace integration	$99.99/mo Google AI Ultra
Claude Cowork	Desktop, app-driving, design + code	Your Mac/Windows desktop	Drives local apps, sees your screen	$20/mo Claude Pro
ChatGPT Codex Mobile	Coding agent on phone	OpenAI cloud + iOS/Android	Approve diffs and redirect work from phone	Included in ChatGPT plans
Grok Agentic (Grok 4.6)	Real-time research, X scraping	SpaceXAI cloud	Native X integration; GDPval-AA v2 Elo 1755	$30/mo SuperGrok
OpenAI Operator-class	Browser tasks, web forms	OpenAI cloud + your browser	Web automation	ChatGPT Pro

Runner-up and alternatives: Claude Cowork is the runner-up overall and the natural pick when you want the agent on your machine driving your apps. ChatGPT Codex Mobile is the runner-up for coding agents. Grok Agentic is the niche pick for real-time research.

What changed this month

No new consumer agents shipped, and Meta's Muse Code (August 5) is a terminal developer tool rather than a consumer one, so the Gemini Spark (cloud) versus Claude Cowork (desktop) choice still drives most agent decisions for individual users. The model layer underneath them moved again: Claude Fable 5.1 (September 1) now leads Artificial Analysis's Agentic Index at 61, ahead of Claude Opus 5 at 59, and tops the GDPval-AA v2 board at 1853 against Opus 5's 1824. Opus 5 is still the one most teams should build on at $5 / $25, half Fable 5.1's price, and it tops Arena's Agent board, with Kimi K3 in third. For teams building their own agents, Meta's Muse Spark 1.2 (August 5) is a cheap agent-native option at $1.25 / $4.25 whose GDPval-AA v2 agentic Elo jumped from 1371 to 1631, though Scale SEAL has rated only the older 1.1, which leads its MCP Atlas tool-use board at 88.1. GLM 5.3 Flash (August 26, MIT) is now the open-weight agent model we would host, at Intelligence Index 57 and AutomationBench 48.8 on Z.ai's own chart, where Claude Opus 4.8 scores 41.0, though it has no Arena votes yet; GLM-5.2 keeps the independent record, at #5 on Arena's Agent board. Grok 4.6 (August 12) is the new cost story in this category rather than a new agent product: it posts a GDPval-AA v2 Elo of 1755, which the September entries pushed down to fifth among distinct models, and Artificial Analysis measures it finishing long-horizon jobs in roughly 53 turns and 0.5 billion input tokens against about 103 turns and 2.0 billion for Opus 5, at $0.84 per index task.

Fello AI running on Mac, iPhone and iPad

All the top AI models, one app

The leading models like ChatGPT, Claude and Gemini, together on Mac, iPhone and iPad.

Free to start, 4.7★ across 27,000+ reviews.

Download Fello AI

Use case guide, September 2026

Best AI for Students

GPT-5.5 Free

Essays & research

→

Gemini 3.6 Flash

STEM + PDFs

→

Claude Sonnet 5

Writing & editing

The best AI for students is GPT-5.5 Free inside ChatGPT for general coursework and Gemini 3.6 Flash Free inside the Gemini app for STEM and multimodal study, with Qwen 3.7 Max as the API alternative for harder problem sets (200 free requests a day) and Claude Opus 5 as the alternative for essay editing. Most students don't need to pay: the free ChatGPT tier now defaults to GPT-5.6 (with GPT-5.5 still available), Gemini 3.6 Flash is in the free Gemini app and AI Studio, Claude Sonnet 5 is the new free Claude default, and DeepSeek V4 is free on DeepSeek's chat site. For step-by-step working on the hardest math, GPT-5.6 Sol is OpenAI's new flagship (its FrontierMath score is not yet published, with GPT-5.5 Pro's verified 39.6% on FrontierMath Tier 4 the current mark), though both are paid-only; Qwen 3.7 Max is the value alternative at 97.1 HMMT 2026 February with API pricing at $1.25 / $3.75 on its current 50% promo ($2.50 / $7.50 list).

Task	Best Model	Why	Free?	Alternative
Essays & coursework	GPT-5.5	Free in ChatGPT, improved factual reliability vs 5.4	Yes	Claude Sonnet 5 (free Claude)
STEM problem-solving	GPT-5.6 Sol / Qwen 3.7 Max	New STEM flagship (5.5 Pro: 39.6% FrontierMath) / 97.1 HMMT 2026 Feb	Pro paid / Qwen API paid	Gemini 3.6 Flash (free)
Research & accuracy	Gemini 3.1 Pro	98% ARC-AGI-1 at $0.52/task, native Google Search grounding	Yes (Gemini app)	Claude Opus 5
Writing editing	Claude Sonnet 5	Free and default on claude.ai; Claude Fable 5 is the quality leader	Yes (Claude free)	Claude Fable 5
Multimodal study (PDFs, slides, images)	Gemini 3.6 Flash	1M context, free in Gemini app	Yes	NotebookLM (Google)

Runner-up and alternatives: Claude Sonnet 5 (free) is the runner-up for essay writing and editing. Gemini 3.6 Flash (free) is the runner-up for multimodal study and PDF ingestion. DeepSeek V4 is the runner-up for problem-solving on a strict zero-cost budget.

Use case guide, September 2026

Best AI for Work & Professionals

GPT-5.6

Daily knowledge work

→

Claude Opus 5

Coding & writing

→

Gemini 3.1 Pro

Research & briefings

The best AI for professional work is GPT-5.6 (ChatGPT's default since July 9) for daily knowledge work, Claude Opus 5 for coding and high-stakes writing, and Gemini Spark for 24/7 agentic workflows. Most professionals get the most out of running two paid subscriptions (ChatGPT Plus at $20/month plus Claude Pro at $20/month, total $40/month), or consolidating with Fello AI at $9.99/month for all five top models in one Mac/iOS app. For agentic work that runs while you sleep, Gemini Spark on Google AI Ultra at $99.99/month is the only true 24/7 cloud agent.

Use Case	Best Model	Key Stat	Price	Alternative
Daily knowledge work	GPT-5.6	ChatGPT's default since July 9; the assistant most people can open	$20/mo ChatGPT Plus	Claude Opus 5
Coding (proprietary)	Claude Opus 5	#1 on Arena WebDev (1,702.9) and image-to-WebDev (1,668.6); Anthropic's recommended default	$20/mo Claude Pro	Claude Fable 5
Coding (cost-effective)	Qwen 3.7 Max	80.4 SWE-Verified, 1M context	$1.25 / $3.75 promo; $2.50 / $7.50 list	DeepSeek V4-Flash 0731
Research & briefings	Gemini 3.1 Pro	98% ARC-AGI-1 at $0.52/task, Google grounding	Google AI Pro / Ultra	Claude Opus 5
Hard math, physics, finance modelling	GPT-5.6 Sol	OpenAI's new STEM flagship (5.5 Pro verified at 39.6% FrontierMath)	$100/mo ChatGPT Pro	Qwen 3.7 Max
Always-on agent workflows	Gemini Spark	First 24/7 cloud agent	$99.99/mo Google AI Ultra	Claude Cowork
Live news, X-context creative	Grok 4.6	Intelligence Index 61 + native X grounding	$30/mo SuperGrok	Gemini 3.1 Pro
All-in-one consolidation	Fello AI	ChatGPT + Claude + Gemini + Grok + DeepSeek	$9.99/mo	Pay each vendor separately

Runner-up and alternatives: for most professional teams, Claude Opus 5 is the runner-up to GPT-5.6 for daily work and now the coding leader outright at $5 / $25, with Claude Fable 5 the runner-up for the hardest long-horizon work. Gemini 3.1 Pro is the runner-up for research-heavy roles, and Gemini Spark is the unique pick if you can put a cloud agent to work on long tasks.

Pricing Comparison

AI Model Pricing in September 2026

From $0 free tiers to $199.99/month Google AI Ultra. The most consequential price on this table is Claude Opus 5 at $5 / $25, because it is the second-highest scorer on Artificial Analysis's Intelligence Index at half the cost of the two Fable models around it. Claude Fable 5.1, which took the top of that index on September 1, holds the same $10 / $50 list price as Fable 5 but cuts cache reads 75% to $0.25, which is where the saving on long agentic runs actually sits rather than on the sticker. Meta's Muse Spark 1.2 lists at $1.25 / $4.25, with a Contributor tier at $0.10 / $0.20 for anyone willing to let Meta train on their prompts, and Grok 4.6 lists at $2 / $6, with the caveat that a prompt of 200,000 tokens or more re-bills the entire request at $4 / $12. The price-performance pick moved in August, and not because anything got cheaper: DeepSeek raised both V4 models roughly fourfold on August 16, taking V4-Flash 0731 from a flat $0.14 / $0.28 to $0.22 / $0.66 off-peak and $0.44 / $1.32 at peak. The pick is now GLM 5.3 Flash at $0.15 / $0.50 list, halved until September 9, which scores 57 on the Intelligence Index and undercuts even DeepSeek's off-peak rate on both sides at every hour of the day. At the frontier itself the value pick is Grok 4.6, scoring 61 for $0.84 per index task where Opus 5 costs $2.34 and GPT-5.6 Sol charges $4 / $20 for the same score after its August 21 cut. Among closed models GPT-5.6 Luna is the cheapest at $0.20 / $1.20 after OpenAI's July 30 cut. For a deeper breakdown see our full AI Pricing Comparison Guide.

Model	Input (per 1M)	Output (per 1M)	Context	Free Access?
GPT-5.5	$5.00	$30.00	1M (400K in Codex)	ChatGPT Free; API paid
GPT-5.5 Pro	$30.00	$180.00	1M	ChatGPT Pro from $100/mo ($200 higher-usage tier)
GPT-5.6 Sol	$4.00	$20.00	Not published	Live in ChatGPT, Codex & API (July 9)
GPT-5.6 Terra	$2.00	$12.00	Not published	Live in ChatGPT, Codex & API (July 9)
GPT-5.6 Luna	$0.20	$1.20	Not published	Live in ChatGPT, Codex & API (July 9)
Claude Opus 5	$5.00	$25.00	1M	Claude Pro/Max default; API paid
Claude Opus 4.8	$5.00	$25.00	1M	Legacy model at Anthropic; Pro/Max/API
Claude Fable 5.1	$10.00	$50.00	1M	Cache reads $0.25; in Max/Team Premium (~50% of usage limits); Pro/Team Standard via credits. Mythos 5.1 bills identically, by invitation only
Claude Fable 5	$10.00	$50.00	1M	Permanent in Max/Team Premium (~50% of usage limits); Pro/Team Standard via credits
Claude Sonnet 5	$2.00	$10.00	1M	Claude Free & Pro default; API paid
Claude Sonnet 4.6	$3.00	$15.00	1M	API paid (superseded by Sonnet 5)
Gemini 3.1 Pro	$2.00 (≤200K) / $4.00 (>200K)	$12.00 (≤200K) / $18.00 (>200K)	1M	Limited Gemini app; API paid
Gemini 3.8 Flash	$0.75 intro / $1.50 from Jan 1, 2027	$3.75 intro / $7.50 from Jan 1, 2027	1M	AI Studio, Antigravity, Gemini Enterprise + paid API; in the Gemini app reported for AI Pro/Ultra
Gemini 3.7 Flash	$0.75 intro / $1.50 from Jan 1, 2027	$3.75 intro / $7.50 from Jan 1, 2027	1M	AI Studio, Android Studio, Antigravity + paid API; in the Gemini app via Spark only (AI Pro/Ultra, excludes EEA/UK/CH/Nigeria)
Gemini 3.6 Flash	$0.75 intro / $1.50 from Jan 1, 2027	$3.75 intro / $7.50 from Jan 1, 2027	1M	Free Gemini app default; AI Studio; free API tier + paid API
Gemini 3.5 Flash-Lite	$0.30	$2.50	1M	AI Studio; free API tier + paid API
Qwen 3.7 Max	$1.25 promo / $2.50 list	$3.75 promo / $7.50 list	1M	200 free requests/day; API paid beyond that
MiniMax M3	$0.30 (50% off $0.60)	$1.20 (≤512K)	1M	Open weights; hosting costs apply
LongCat-2.0	Provider-dependent	Provider-dependent	1M	Open weights (MIT); hosting costs apply
NVIDIA Nemotron 3 Ultra	Provider-dependent	Provider-dependent	1M	Open weights (OpenMDW); hosting costs apply
Qwen 3.5 (open-weight)	Self-host / Together	Self-host / Together	1M	Open weights; hosting costs apply
Nex-N2-Pro	Self-host / providers	Self-host / providers	1M	Open weights (Apache 2.0); hosting costs apply
Rio 3.5 Open 397B	Self-host / providers	Self-host / providers	1M	Open weights (MIT); hosting costs apply
Grok 4.3	$1.25	$2.50	1M	Free consumer plan; API paid
Grok 4.6	$2.00 (<200K) / $4.00 (≥200K)	$6.00 (<200K) / $12.00 (≥200K)	500K	SpaceXAI API, Cursor, Grok Build, OpenRouter, Vercel, Cloudflare
Muse Spark 1.2	$1.25 ($0.10 Contributor tier)	$4.25 ($0.20 Contributor tier)	1M	Paid API; Muse Code, Meta Model API and OpenRouter; Contributor tier trades training rights for a ~92% discount
Kimi K3	$3.00 ($0.30 cache-hit)	$15.00	1M	Free basic tier in the Kimi app; open weights on Hugging Face (Kimi K3 License)
Gemini Omni Flash (video)	$1.50	$17.50 (video output)	10-second clips	Gemini app / Flow; AI Studio + API
DeepSeek V4-Pro	$0.66 off-peak / $1.32 peak ($0.022 cache-hit off-peak)	$1.98 off-peak / $3.96 peak	1M	DeepSeek Chat free; API paid, peak/off-peak tiers since August 16
DeepSeek V4-Flash 0731	$0.22 off-peak / $0.44 peak ($0.007 cache-hit off-peak)	$0.66 off-peak / $1.32 peak	1M	DeepSeek Chat free; API paid, peak/off-peak tiers since August 16
Kimi K2.7 Code	Provider-dependent	Provider-dependent	256K	Open weights; hosting costs apply
GLM-5.2	Provider-dependent	Provider-dependent	1M	Open weights; hosting costs apply
GLM 5.3	$1.40 ($0.26 cache-hit)	$4.40	1M	Z.ai API, GLM Coding Plan and ZCode; weights on Hugging Face since August 28 under the bespoke GLM 5.3 License
GLM 5.3 Flash	$0.15 ($0.03 cache-hit); $0.075 promo	$0.50; $0.25 promo	1M	Z.ai API, GLM Coding Plan, OpenRouter; MIT weights on Hugging Face; promo ends September 9
Tencent Hy4 Preview	$0.834 ($0.042 cache-hit)	$2.501	1M+	Open weights (Apache 2.0); Tencent Cloud TokenHub, OpenRouter, WorkBuddy, CodeBuddy
Qwen3.8-Flash-Next	Provider-dependent	Provider-dependent	262K (to 1M)	Open weights (Qwen Community License 1.0); hosting costs apply
ERNIE 5.1	China-region pricing	China-region pricing	256K	Baidu free tier
Gemini Spark (agent)	Not API-priced	Not API-priced	1M (Gemini base)	Google AI Ultra $99.99 or $199.99/mo
Fello AI (aggregator)	Routed via app	Routed via app	Model-dependent	$9.99/mo, free tier available

The GPT-5.5 and GPT-5.5 Pro rates above are short-context prices; OpenAI no longer publishes the specific long-context figures. The GPT-5.6 tiers are billed at 2x input and 1.5x output once a prompt passes 272K input tokens, which puts long-context Terra at $4 / $18 and Luna at $0.40 / $1.80. Grok 4.6 works the same way but bites harder: at 200,000 prompt tokens and above, SpaceXAI re-bills the entire request at the higher rate rather than only the overage, so crossing the line by a thousand tokens doubles the cost of the whole call. DeepSeek is the other rate card that needs reading twice: since August 16 both V4 models bill at peak rates from 01:00 to 04:00 and from 06:00 to 10:00 UTC on weekdays, and at exactly half that in every other hour, so the same job can cost twice as much depending on when you run it. If you want access to multiple AI models without managing separate subscriptions, Fello AI provides GPT, Claude, Gemini, Grok, Perplexity, and more in a single app for Mac, iPhone, and iPad from $9.99/month.

Open-Weight Models

Best Open-Weight Models in September 2026

The best open-weight model in September 2026 is Kimi K3, and it took the lead the moment Moonshot published the weights on July 27, 2026. It holds the highest Intelligence Index of any open model at 60 on Artificial Analysis, seven points clear of GLM-5.2, and on Arena it is still the highest-placed open entry, at #2 on WebDev (1,675.5) behind only Claude Opus 5 and #3 on the Agent board. One catch decides which of the two you should actually use. The K3 download is 96 safetensors shards and about 1.56 TB, which needs a multi-node GPU cluster rather than a workstation, and it ships under a custom Kimi K3 License rather than MIT. So the practical recommendation for teams running their own weights is now GLM 5.3 Flash (Z.ai, MIT), released on August 26 at Intelligence Index 57, four points above GLM-5.2 and at less than half the size, 320B total with 18B active. It is the first natively multimodal model in the GLM-5 line, it takes a 1M-token context, and the weights were on Hugging Face under MIT the day it launched. GLM-5.2 (MIT) is the fallback, at Intelligence Index 53, and it still holds the independent record the newer model has not had time to earn, #6 on Arena WebDev (1,587.1) and #5 on the Agent board. Third place changed hands on the last day of July: DeepSeek V4-Flash 0731 was re-post-trained and jumped from Intelligence Index 40 to 52 on the current v4.1.1 index under MIT, though its API price has since gone up roughly fourfold, to $0.22 / $0.66 off-peak and $0.44 / $1.32 at peak. Three releases then closed the month, and none of them has an independent rating yet. GLM 5.3 finally published its weights on August 28, on exactly the date its placeholder had carried, but under a bespoke GLM 5.3 License rather than MIT, with a clause requiring any model-as-a-service operator above $10 billion in revenue to pass a Z.ai security review first; note also that GLM 5.3 Flash is not a trimmed GLM 5.3 but a separate model on a newly trained base, which is why it could ship under plain MIT while the flagship could not. Tencent open-sourced Hy4 Preview on August 28 at 770B total and 49B active under Apache 2.0, the largest permissively licensed model published so far, on the strength of an internal blind evaluation rather than any independent board. And Alibaba published Qwen3.8-Flash-Next on August 26, a 125B mixture-of-experts with only 6B active that previews the Qwen4 architecture, under the Qwen Community License 1.0 rather than Apache. All three are listed below rather than ranked.

Model	Best For	Key Benchmark	Context / License	Where To Run
Kimi K3	Highest-scoring open model, agentic and web work	II 60, highest of any open model; #2 Arena WebDev, #3 Arena Agent; 2.8T/104B active	1M / Kimi K3 License	Hugging Face (96 shards, 1.56 TB), Moonshot API, providers
GLM 5.3 Flash	Best open model most teams can actually run	II 57 (v4.1.1), on the intelligence-versus-cost Pareto frontier at about $0.09 an index task; 320B/18B active, natively multimodal	1M / MIT	Hugging Face, Z.ai API ($0.15/$0.50), OpenRouter
GLM-5.2	Fallback with an independent track record	II 53; #6 Arena WebDev (1,587.1), #5 Arena Agent; 744B/40B active	1M / MIT	Z.ai, Hugging Face, OpenRouter
GLM 5.3	Open since August 28, but on a bespoke licence	Same 744B base as GLM-5.2, post-trained only, published checkpoint counted at 753B; CyberGym 84.5% and Terminal-Bench 3.0 28.3 (vendor figures)	1M / GLM 5.3 License (MaaS above $10B revenue needs a Z.ai security review)	Hugging Face, Z.ai API ($1.40/$4.40), GLM Coding Plan, ZCode
DeepSeek V4-Flash 0731	Best open value if you host it yourself	II 52 (v4.1.1), up from 40 on July 31; 284B/13B active	1M / MIT	DeepSeek API ($0.22/$0.66 off-peak, $0.44/$1.32 peak since August 16), local
Tencent Hy4 Preview	Largest permissively licensed model yet	770B/49B active; 2.99/4.00 on Tencent's own 203-task expert panel vs Kimi K3 2.94 and GLM 5.3 2.92 (vendor); no independent rating yet	1M+ / Apache 2.0	Hugging Face (BF16 and FP8), Tencent Cloud TokenHub, OpenRouter
Qwen3.8-Flash-Next	Preview of the Qwen4 architecture	125B total plus a 51B N-gram embedding, 6B active; 91.7 GPQA Diamond and 62.5 SWE-Bench Pro (vendor); no independent rating yet	262K to 1M / Qwen Community License 1.0	Hugging Face, providers, self-host
LongCat-2.0	Frontier open coder trained on Chinese chips	II 33 (v4.1); 59.5% SWE-Bench Pro (vendor), 1.6T/~48B active	1M / MIT	Hugging Face, GitHub, OpenRouter
MiniMax M3	Cheap frontier-class multimodal	II 44 (v4.1), 59% SWE-Bench Pro, multimodal	1M / license TBD	Hugging Face, API $0.30/1M (50% off)
Nex-N2-Pro	Strongest open coding score	II 41 (v4.1); 80.8 SWE-Bench Verified, 397B/17B active	Qwen-based / Apache 2.0	Hugging Face, providers, self-host
Kimi K2.7 Code	Strongest commercially-licensed open coder	+21.8% on Kimi Code Bench v2 vs K2.6 (vendor); 1T/32B active	256K / Modified MIT	Hugging Face, DeepInfra, providers
DeepSeek V4-Pro	Agentic real-world work	II 44 (v4.1), 1.6T/49B active	1M / MIT	DeepSeek API ($0.435/$0.87), local
Hy3	Newest permissive-licence entrant	II 41 (v4.1); #22 on Arena WebDev (1,516.4)	Apache 2.0	Hugging Face, providers, self-host
Inkling	Thinking Machines' first open model	II 41 (v4.1), agentic 32.3, released July 15	Open weights	Hugging Face, providers, self-host
Inkling Small	Same family at a quarter the size	II 40 (v4.1), agentic 30.8; 276B/12B active, text, image and audio in	Apache 2.0	Hugging Face (BF16 and NVFP4), providers, self-host
NVIDIA Nemotron 3 Ultra	NVIDIA-tuned, fully permissive license	II 38 (v4.1), 65-70.4 SWE-Bench Verified, 550B/55B active	1M / OpenMDW	OpenRouter, Hugging Face, AWS (8x B200 self-host)
Qwen 3.5 (397B / 17B active)	Multimodal, fast decode	88.4 GPQA, 91.3 AIME 2026, 83.6 LiveCodeBench v6	1M / open	Together, OpenRouter, local
Qwen3.6-35B-A3B	Efficient open agentic coder (3B active)	86.0 GPQA Diamond, 92.7 AIME 2026, 35B/3B active	262K (to 1M YaRN) / Apache 2.0	Hugging Face, OpenRouter, local
Qwen3.6-27B	Laptop-runnable dense coder	87.8 GPQA Diamond, dense 27B, multimodal	256K / Apache 2.0	Local Mac/PC, Hugging Face, OpenRouter
Rio 3.5 Open 397B	Qwen 3.5 fine-tune, multilingual reasoning	70.8 Terminal-Bench 2.1 (first-party), beats Qwen 3.7 Plus on 4/5	397B/17B active / MIT	Hugging Face, providers, self-host
Llama 4 Maverick	Meta-line flagship	17B active / 400B total params	Llama 4 license	Meta cloud, Hugging Face, local
NVIDIA Nemotron 3 Nano Omni	Edge / low-power	Multimodal, very small footprint	Compact / open	Local, NVIDIA tool

Licensing matters as much as raw score here, and August widened the gap between the two groups. Kimi K2.7 (Modified MIT), DeepSeek V4 (MIT), GLM 5.3 Flash (MIT), GLM-5.2 (MIT), LongCat-2.0 (MIT), Hy3 (Apache 2.0), Hy4 Preview (Apache 2.0), Nex-N2-Pro (Apache 2.0) and Nemotron 3 Ultra (OpenMDW) all clearly allow commercial use with no revenue test. The rest carry conditions worth reading before you build on them: MiniMax M3 and MiniMax H3 ship under their own community licences, H3 additionally requiring an application from the US, EU, UK and South Korea; Kimi K3 sits on a custom licence with a revenue threshold for anyone reselling it as a service; GLM 5.3 requires a Z.ai security review of any model-as-a-service operator above $10 billion in revenue; and the Qwen Community License 1.0 on Qwen3.8-Flash-Next requires a separate licence from Qwen to run a model-as-a-service or an AI work-assistant business, though internal use is exempt. No open model is first on any Arena board we track any more: Claude Opus 5 took the Agent board in July, the last one an open model led.

How We Evaluate

Benchmarks, Prices, and Hands-On Use

Every ranking on this page combines three inputs: public benchmarks from seven independent houses (Artificial Analysis, Arena formerly LMArena, Scale SEAL, LiveBench, EQ-Bench, ARC Prize and the official Terminal-Bench 2.1 board, covering the Intelligence and Agentic indexes, GPQA Diamond, ARC-AGI-1 through 3, Humanity's Last Exam, GDPval-AA, FrontierMath, HMMT, MCP Atlas, SWE Atlas and the Remote Labor Index), published API and subscription pricing from each vendor's official pricing page, and hands-on use by the FelloAI editorial team running real prompts across the same task on every model. We re-fetch official pricing and benchmark sources before every monthly update.

Benchmarks are weighted to the use case: SWE-bench and Terminal-Bench drive coding, GPQA Diamond and ARC-AGI-2 drive accuracy, GDPval-AA (Artificial Analysis's professional-deliverables benchmark) informs professional-task quality while writing style is judged primarily by hands-on testing, FrontierMath and HMMT drive problem-solving. We disclose when a benchmark is vendor-reported but not independently verified, and we strip any claim we cannot reproduce against a live source. We do not move a category crown on the strength of a single leaderboard, and where a new model is too recent for the human-preference boards to have rated it, we say so rather than crowning it on benchmark scores alone. When a model goes through a major upgrade between updates, we re-rank the category and add a "What changed this month" line at the bottom of the deep-dive.

Not sure which model to pick?

Fello AI brings the top AI models together in one lightweight native app.

Switch between them anytime, no separate subscriptions.

Download Fello AI

Frequently Asked Questions

Common Questions

What is the best AI model right now in September 2026?

It depends on the task. On overall benchmark score, Claude Fable 5.1 (September 1) is #1 on Artificial Analysis's Intelligence Index at 66 and its Agentic Index at 61, with Claude Opus 5 second on both at 63 and 59; Arena has rated Opus 5 top of four of its boards, including both coding boards, and has no votes for Fable 5.1 yet. Grok 4.6 (August 12) is the value pick at the frontier, tying GPT-5.6 Sol at 61 for a third of the price. For daily chat, GPT-5.6 has been ChatGPT's default since July 9 and is the assistant most people can actually open, although Claude Fable 5 leads Arena's text leaderboard. For coding, Claude Opus 5 is #1 on Arena's WebDev (1,702.9) and image-to-WebDev (1,668.6) boards at half Fable 5's price. For writing, Claude Fable 5 is #1 on Arena's text leaderboard (1508.6), with Fable 5.1 ahead of it on Humanity's Last Exam (59.1%) and level with it on AA-Omniscience (43). For accuracy, Gemini 3.1 Pro ties the human panel on ARC-AGI-1 at 98% for $0.52 a task. For hard math and reasoning, GPT-5.6 Sol is #1 on LiveBench Mathematics, LiveBench Reasoning and ARC-AGI-2. For images, ChatGPT Images 2.0 leads both boards, and for video Gemini Omni Flash leads both. For agents, Gemini Spark is the 24/7 cloud agent and Claude Cowork the desktop one.

What is Claude Opus 5?

Claude Opus 5 is Anthropic's July 24, 2026 flagship, and it was the #1 model on Artificial Analysis until Claude Fable 5.1 arrived on September 1. It now sits second on the Intelligence Index at 63 against 66, second on the Agentic Index at 59 against 61, and second on the GDPval-AA v2 professional-deliverables board at 1824 against 1853, still clear of Grok 4.6's 1755 and Claude Fable 5's 1723. API pricing is $5 / $25 per 1M tokens, the same as Opus 4.8 and half the cost of Fable 5, with a 1M-token context window and a May 2026 knowledge cutoff. ARC Prize independently confirms Anthropic's launch claim on ARC-AGI-3, where Opus 5 scores 30% against 8% for the next-best model, roughly 3.75x. Arena has since rated it, placing it #1 on WebDev, image-to-WebDev, Document and the Agent board and #6 on text, and that is what moved the coding crown to it. One thing to keep in mind: Artificial Analysis measures its hallucination rate at 50%, which is why it holds no accuracy crown on this page.

What is Claude Fable 5.1?

Claude Fable 5.1 is Anthropic's September 1, 2026 release and the highest-scoring model Artificial Analysis has measured, at 66 on its Intelligence Index at max effort and 61 on its Agentic Index. It shipped alongside Claude Mythos 5.1, which is the same model with the biology and cybersecurity safeguards relaxed, available by invitation only and with no published eligibility criteria. Pricing is $10 / $50 per 1M tokens, unchanged from Fable 5, but cache reads fall 75% to $0.25, on a 1M-token context window with a June 2026 knowledge cutoff. It is included in Claude Max and Team Premium at roughly 50% of weekly limits and reachable from Pro through usage credits. Anthropic still recommends starting with Claude Opus 5 for most workloads, since it costs half as much and runs faster. Our full Claude Fable 5.1 and Mythos 5.1 breakdown covers the system card and the safeguard split.

Is Claude Fable 5 back?

Yes. Anthropic redeployed Claude Fable 5 on July 1, 2026 after the US government lifted the export-control restriction it had imposed on June 12. It is available again on the Claude API, Claude.ai, Claude Code, and Claude Cowork. Anthropic made Fable 5 permanent in the paid plans on July 20, 2026. Max and Team Premium include it at roughly 50% of regular usage limits; Pro and Team Standard reach it through usage credits with a one-time $100 starting credit. API pricing is $10 / $50 per million tokens. Fable 5 is a Mythos-class model built for long-horizon agentic work with a 1M-token context, and it is the runner-up for coding behind Claude Opus 5, at #2 on Arena's image-to-WebDev board (1,625.7) and #4 on WebDev.

What is GPT-5.6 and can I use it?

Yes, as of July 9, 2026. GPT-5.6 is OpenAI's next-generation model family, and after a two-week gated preview that began June 26 behind a US-government safety review, it reached general availability on July 9. There are three tiers, least to most capable: Luna, the fast, cheapest tier ($0.20 / $1.20 per 1M tokens); Terra, a balanced everyday model OpenAI says matches GPT-5.5 ($2 / $12); and Sol, the flagship tuned for biology, chemistry, and cybersecurity ($4 / $20). OpenAI cut Luna by 80% and Terra by 20% on July 30, 2026 and left Sol alone, then cut Sol by more than 20% on August 21, 2026 as a promotional rate for about three months. No ChatGPT subscription price has changed. It is now live across ChatGPT, Codex, and the API as OpenAI's default model. One caveat: OpenAI's system card and the external evaluator METR flagged elevated "scheming" behaviour in Sol, so treat it carefully for high-stakes factual work until independent results settle.

Is Grok 4.6 out yet?

Yes. SpaceXAI released Grok 4.6 on August 12, 2026, replacing Grok 4.5 as the flagship just over a month after it shipped. It is not a new foundation model: the company kept the Grok 4.5 base and improved it through additional post-training, including reinforcement learning in agentic environments, and published no new architecture or parameter count. Artificial Analysis scores it at Intelligence Index 61, up five points from Grok 4.5's 56, which ties GPT-5.6 Sol at 61, behind Claude Fable 5.1 at 66, Claude Opus 5 at 63 and Claude Fable 5 at 62. Pricing is unchanged at $2 / $6 per 1M tokens on a 500K-token context window, though prompts of 200,000 tokens and above re-bill the whole request at $4 / $12. It is live in the SpaceXAI API, Cursor, Grok Build, OpenRouter, Vercel and Cloudflare. Grok 4.6 is also our creativity pick, on product grounds rather than prose quality; for the best-written output, Claude Fable 5 wins outright. Full numbers are in our Grok 4.6 breakdown.

Which AI is the best for coding?

Claude Opus 5 is the best for coding, and it wins the two boards where developers vote on the result rather than a harness running a script: #1 on Arena's WebDev board (1,702.9) and #1 on image-to-WebDev (1,668.6), on August 1 and July 31 vote cutoffs. It runs $5 / $25 per 1M tokens, half of Claude Fable 5, and Anthropic's own docs say to start with Opus 5 for complex agentic coding. Claude Fable 5 is the runner-up for the hardest long-horizon work at #2 image-to-WebDev and #4 WebDev, and Kimi K3 is the contender at #2 on WebDev (1,675.5). Note that Artificial Analysis's Coding Index is not a Claude sweep: GPT-5.6 Sol (xhigh) leads it at 78.3 with Opus 5 at 78.0, close enough to call a tie. GLM 5.3 Flash (MIT, August 26) is now both the price-performance pick, at $0.15 / $0.50, and the best open-weight coder you can host yourself, at Intelligence Index 57; it took the first of those from DeepSeek V4-Flash 0731 when DeepSeek raised its rates roughly fourfold on August 16.

Which AI is the best for writing?

Claude Fable 5 is the best for writing, and it is the only model in the top three of all three independent writing boards: #1 on Arena creative writing, #1 on LiveBench Language (90.7) and #3 on EQ-Bench Creative Writing v3. It costs $10 / $50 per 1M tokens. Claude Sonnet 5 is the value pick and what most people should actually use, since it is free and default on claude.ai at $2 / $10, a launch rate Anthropic made permanent in August 2026, though it ranks #53 on Arena creative writing. Kimi K3 wins EQ-Bench outright at 2377 if you want distinctive fiction, Claude Fable 5.1 tops the GDPval-AA v2 professional-deliverables board at 1853 with Claude Opus 5 second at 1824, GPT-5.5 is the alternative for fact-anchored business writing, and Gemini 3.8 Flash is the price-performance pick for bulk content.

What is the best open-weight AI model in 2026?

Kimi K3 is the best open-weight model right now, and it has been since Moonshot published the weights on July 27, 2026. It has the highest Intelligence Index of any open model at 60 on Artificial Analysis, and on Arena it is #2 on WebDev and #3 on the Agent board. The catch is that it is 96 shards and about 1.56 TB under a custom Kimi K3 License, so GLM 5.3 Flash (August 26, MIT) is now the model most teams can actually host, at Intelligence Index 57 with 320B total and 18B active parameters. Third is DeepSeek V4-Flash 0731, which jumped from Intelligence Index 40 to 52 on July 31 without changing size or licence, though DeepSeek then raised its API rates roughly fourfold on August 16, to $0.22 / $0.66 off-peak and $0.44 / $1.32 at peak, which is what cost it the price-performance pick. Note that GLM 5.3, released August 14, is a different and larger model whose weights did land on August 28, but under a bespoke GLM 5.3 License rather than MIT. Two more open models closed the month with no independent rating yet: Tencent's Hy4 Preview on August 28, 770B under Apache 2.0 and the largest permissively licensed model so far, and Alibaba's Qwen3.8-Flash-Next on August 26, a preview of the Qwen4 architecture under the Qwen Community License 1.0. Worth knowing honestly, no open model is first on any Arena board we track any more.

What is the cheapest frontier-class AI model?

On API pricing per million tokens, GPT-5.6 Luna is now the cheapest closed frontier-class model at $0.20 / $1.20, after OpenAI cut it 80% on July 30, 2026, and Artificial Analysis scores it at Intelligence Index 52 on its v4.1.1 index, level with Gemini 3.6 Flash. Cheaper still is GLM 5.3 Flash, our price-performance pick since August, at $0.15 / $0.50 list and half that until September 9, scoring 57 on the Intelligence Index under a plain MIT licence you can self-host. DeepSeek V4-Flash 0731 held that pick until DeepSeek raised its rates roughly fourfold on August 16, to $0.22 / $0.66 off-peak and $0.44 / $1.32 at peak. Google's own cheap tier improved again on September 2, with Gemini 3.8 Flash scoring 59 at the same introductory $0.75 / $3.75, though that rate doubles on January 1, 2027. MiniMax M3 now lists at $0.30 per million input tokens on a permanent 50% discount, with LongCat-2.0 as a strong MIT alternative. Qwen 3.7 Max at $1.25 / $3.75 on its current promo is the value pick at Intelligence Index 46, though Luna now undercuts it on both price and score.

Which AI models are free?

ChatGPT Free now defaults to GPT-5.6 (with GPT-5.5 still available) under usage limits. Gemini Free runs Gemini 3.6 Flash in the Gemini app and Google AI Studio. Claude Free runs Claude Sonnet 5 (the new default) with daily limits. DeepSeek Chat runs DeepSeek V4 free on the DeepSeek website. Grok has a limited free consumer plan (X Premium is a paid add-on). Qwen 3.5, Qwen3.8-Flash-Next, NVIDIA Nemotron 3 Ultra, MiniMax M3, LongCat-2.0, Kimi K2.6, DeepSeek V4, GLM-5.2, GLM 5.3, GLM 5.3 Flash and Tencent's Hy4 Preview are open-weight and free to self-host, though the licences differ and only some of them are plain MIT or Apache. Qwen 3.7 Max is API-only with no consumer chat front-end, but Alibaba now includes 200 free model requests per day. Kimi has a free basic tier in its app, with heavier agentic use metered.

What is Gemini Spark and is it worth $99.99/month?

Gemini Spark is Google's first 24/7 cloud-resident AI agent, launched at Google I/O on May 19, 2026 and exclusive to the Google AI Ultra plan, which Google restructured to include a $99.99/month entry tier and a $199.99/month top tier. Spark is built on Gemini base models with Google's Antigravity harness on a Google Cloud VM, integrates with Gmail, Google Docs, and other Google Workspace apps, and can interact with Chrome and Android's Halo system on the device side. It is worth the spend for users who have repeatable long-running workflows (inbox triage, research roll-ups, scheduled tasks). For one-off tasks, Claude Cowork at $20/month covers most desktop-agent needs.

What is Fello AI?

Fello AI is an AI chatbot for Mac, iPhone, and iPad that lets you use all top AI models like ChatGPT, Claude, Gemini, Grok, and DeepSeek in one app, with models updated regularly so you always have the latest. It is $9.99/month with a 4.7-star rating across 27,000+ reviews.

How often do you update this page?

We update this page at least monthly and within 24-48 hours of any major model launch.