Mistral Large 4 (Le Chonk): Specs, Benchmarks, Price and When the Weights Land

Mistral Large 4, nicknamed Le Chonk, is a 1-trillion-parameter mixture-of-experts model with 49 billion active parameters. Mistral AI released it as a public preview API on October 6, 2026, and says the open weights will follow by the end of October. It reads text and images, answers in text, and costs $1.36 per million input tokens at list price.

Mistral calls it the strongest open-weight model built in the US or Europe, and competitive with the best open models anywhere. The first independent numbers back the first half of that claim and complicate the second. This guide puts Mistral's benchmarks next to the scores Artificial Analysis and Vals.ai published on launch day. It also covers the specs, the price, and what "open weights" means for a model you cannot download yet.

The Key Takeaways

• Size: 1.05 trillion total parameters, 49 billion active at launch (the docs card now says 52 billion), plus a 1.6B vision encoder. Text and image in, text out.
• Independent score: 38 on the Artificial Analysis Intelligence Index v4.3.2, far ahead of US open models like Nemotron 3 Ultra (23), but 6 to 8 points behind GLM-5.3, Kimi K3 and MiMo-V2.6-Pro.
• Price: $1.36 in / $4.18 out per million tokens at list. Mistral's model page currently shows half that, $0.68 / $2.09, without saying why or for how long.
• Weights: not out yet. Mistral says by the end of October; reporters were told October 27. No licence has been named.
• The real pitch: cyber defense. Mistral reports 82% on a reproduce-and-patch vulnerability test where, it says, Claude Opus 5.5 and GPT-6 Astra score near zero because they refuse.

What Mistral Large 4 Is

Mistral Large 4 is Mistral AI's largest and most capable model so far, a sparse mixture-of-experts network that only switches on about 5% of its weights for each token. Mistral's own label for it is an "open-weight hybrid instruct-and-reasoning MoE with multimodal input", which means one model handles quick answers, step-by-step reasoning and agentic tool use.

Mistral announced it on X at 13:06 UTC. The post carries the four claims the launch hangs on: size, the open-weight ranking, the target industries and the European infrastructure.

Meet Mistral Large 4, aka Le Chonk.

• 1T parameters, natively multimodal. 49B active.
It is the best open weights model from US or Europe on aggregated benchmarks.
• State-of-the-art on critical workloads, including cyber defense, manufacturing and finance and it surpasses closed frontier models on visual grounding.
• Forged in Europe end-to-end and is deployable from Europe via our own Mistral Cloud infrastructure.
• Available to all via API today. Working with cybersecurity partners privately.

Open weights release end of October.

— Mistral AI (@MistralAI) October 6, 2026

The specs below come from Mistral's docs model card and announcement. Where an independent tracker lists a different figure, the table says so.

Two specs do not agree across sources. Mistral's announcement and both launch posts on X say 49 billion active parameters, and so did the docs card when we first read it on launch day. Less than an hour later the card read 52 billion, with no note explaining the change. The context window is the bigger gap. Mistral's docs card lists 1M tokens. Artificial Analysis lists 524,288 tokens and Vals lists 512K with a 256K output cap. Until Mistral explains the gap, plan around 512K.

Why it is called Le Chonk

Le Chonk is internet slang for a very fat cat, and the name is a joke the community started. VentureBeat traces it to Le Chaton Fat, a fictional giant Mistral model that went viral on X and Reddit in June with fake benchmark charts. Mistral CEO Arthur Mensch replied at the time that it was "actually le gros chaton". Chief scientist Guillaume Lample told VentureBeat that Mistral enjoyed the meme and that ML4 could be read as a first version of the idea, with larger models to come.

The training run was not a joke. Mistral says it trained ML4 from scratch on 3,800 NVIDIA Grace Blackwell GPUs in its own European data centers, and serves the preview from the same machines. VentureBeat, CNBC and Euronews all report 4,000 GPUs over about two months, which appears to come from the press briefing. Mistral's written announcement says 3,800.

Our Le Chaton Fat explainer ended with a simple test for any trending model: no weights, no API and no docs means it is probably a kitten. Le Chonk passes two of the three today. The weights are the third.

Mistral Large 4 Benchmarks: Mistral's Numbers vs Independent Ones

Mistral Large 4 outscores every US and European open model on the independent data published so far, but it is not level with the best Chinese open models. Mistral's own charts and the first third-party scores tell two slightly different stories, so they are separated here.

What Mistral reports

Mistral reports 61.7% on DeepSWE v1.1 for coding, the headline number of its launch announcement. It adds 59.4% on SWE-Atlas-QnA and 28.3% on Terminal-Bench 4.0, for a combined Coding Agent Index of 49.8%. Mistral says that puts it ahead of DeepSeek V4 Pro 0813 and Qwen3.8 Max.

VentureBeat checked the coding chart against the live DeepSWE leaderboard. That board, which picks each model's best published agent setup, puts GLM-5.3 and Kimi K3 at about 69%, and GPT-6 Astra, Gemini 3.8 Flash and Claude Opus 5 around 74%.

So 61.7% is competitive, not a lead.

The other headline claims are all Mistral-reported. Mistral Large 4 scores 59.9% on AutomationBench, a set of 657 business workflows across Gmail, Google Sheets, Slack and Salesforce. It reaches 1,393 Elo on AA-Briefcase for long-horizon knowledge work. On the Dense 200 visual grounding test it scores 42% against GPT-6 Astra's 41%, the basis for the launch post's line that it "surpasses closed frontier models on visual grounding". In a blind coding-quality test run with Surge AI, annotators rated it 3.74 out of 5, second of five models and behind only Claude Opus 5 at 4.22.

What independent evaluators measured

Artificial Analysis gives Mistral Large 4 Preview a score of 38 on its Intelligence Index v4.3.2, a composite of 10 evaluations. That is far above US open models on AA's board, such as NVIDIA's Nemotron 3 Ultra (23), Google's Gemma 4 31B (15) and OpenAI's gpt-oss-120b (12). It is also clearly below the top Chinese open models, each measured on its own AA model page on launch day.

Two more details on the AA page matter. Mistral Large 4 is fast, at about 116 output tokens per second, but very verbose: it produced 200 million tokens running the index, against a median of 81 million. That verbosity is why AA puts its cost per index task at $1.13 despite the modest list price. Our guide to AI benchmarks explains what the index measures and why a single number never settles a comparison.

Vals.ai, the second independent evaluator, scored it 48.05% on the Vals Index, 32nd of 44 models overall and ninth among open-weight models. Its standout result is Harvey's Legal Agent Benchmark, where it scores 15.83% and ranks sixth of 75, first among open models. Vals posted that ranking within minutes of the launch.

Mistral Large 4 is now the #1 open-weight model on HLAB and #9 among open-weight models on the Vals Index.

— Vals AI (@ValsAI) October 6, 2026

The honest summary: the claim that ML4 is the best open model from the US or Europe holds on Artificial Analysis. The claim that it competes with the strongest open models globally is a stretch. On Artificial Analysis, GLM 5.3 and Kimi K3 lead it by six to seven points. Mistral says the reinforcement learning run is still going, so the final checkpoint may close some of that gap.

The Cyber Pitch Is the Real Differentiator

Mistral Large 4's strongest argument is cybersecurity, not general intelligence. Mistral says it ranks in the top five models worldwide on the Artificial Analysis Cyber Index and leads open-weight models built outside China by a wide margin. One of the index's tests asks a model to reproduce a real vulnerability in open-source software and then patch it. Mistral reports 82% on it, the highest of any model. It also says ML4 solves 93% of the 40 challenges in Cybench.

The interesting part is why it wins that test.

According to Mistral, Claude Opus 5.5 and GPT-6 Astra score near zero on the same task because they refuse to perform it. Mistral argues that defenders often need to prove a flaw is real before fixing it. A provider's safety filter can block exactly that work in the middle of an incident. An open-weight model running on your own servers cannot be switched off by a policy change.

Mistral is also hedging that pitch. Until the weights ship, it is red-teaming the model with cybersecurity firms, vetted partners and state authorities, who get access with reduced moderation and expanded cyber capabilities. Mistral also reports that ML4 refuses malicious cyber prompts more often than any other open model it tested. Lample framed the goal on launch day as letting enterprises and governments "defend themselves against threat actors that are jailbreaking closed models to perform cyber attacks", as quoted by CNBC.

Z.ai faced the same tension in August. It launched GLM 5.3 without weights and said on X they would be released "in stages following rigorous safety evaluations". Mistral has made a firmer promise, and October 27 is now the test of it.

Is Mistral Large 4 Open Source Yet?

No. Mistral Large 4 is announced as an open-weight model, but on launch day only the preview API exists. Mistral's announcement says the weights will be released "by the end of the month", together with details on the architecture, more benchmarks and the post-training method. VentureBeat and Euronews were given a specific date, October 27.

Lample's launch thread explains the gap. The reinforcement learning run behind the preview is still in progress, and Mistral plans to ship a final version alongside the weights.

Today, we are launching a preview of our new model, Mistral Large 4 (ML4), aka le Chonk 🐈.
ML4 is a 1T-parameter model with 49B active parameters, trained natively with multimodal capabilities.

It is at the frontier of open models, and by far the strongest open-weight model from the US or Europe.
The RL run behind this preview is still in flight and shows no sign of saturation -- we will release a final version before the end of the month along with the weights of the model.
🧵
1/n

— Guillaume Lample (@GuillaumeLample) October 6, 2026

The licence is the open question. Mistral's docs card marks the model "Open" but names no licence, and VentureBeat reports the weights are expected under a custom Mistral licence. That matters, because Mistral's lineup is not uniformly Apache 2.0. Large 3 ships under Apache 2.0, while Mistral Medium 3.5 ships under a modified MIT licence, according to Mistral's own docs. Until Mistral publishes terms for Large 4, do not assume free commercial use.

Artificial Analysis currently files Mistral Large 4 Preview under "proprietary", and nothing is downloadable yet. If you need weights you can run today, the best open source AI models list covers what is actually available.

Mistral Large 4 Pricing

Mistral Large 4 lists at $1.36 per million input tokens and $4.18 per million output tokens, with cached input at $0.14. Those are the figures in Mistral's announcement and on both the Artificial Analysis and Vals pages.

Mistral's docs model card shows something different. It displays those three prices struck through and replaced with exactly half: $0.68 input, $0.07 cached input and $2.09 output. Neither the card nor the announcement says whether this is a preview discount, how long it lasts, or what happens when the final model ships. If you are budgeting, use the list price and treat the half price as a bonus.

Mistral Large 4 costs more than Large 3, which stays on sale at $0.50 and $1.50 for teams that want the older open model. Our Mistral AI pricing guide covers the Vibe subscriptions and the rest of the API lineup. Mistral benchmarks against DeepSeek V4 Pro more than any other model, and our DeepSeek V4 breakdown covers that rival.

How to Try Mistral Large 4 Today

The only way to use Mistral Large 4 on launch day is the preview API in Mistral Studio, Mistral's developer platform. Create a Studio account, add billing, generate an API key and call the model by name from the chat completions or conversations endpoints. The docs card lists structured outputs, function calling, document Q&A, batching and Mistral's built-in agent tools as supported.

Mistral's announcement does not mention Vibe, the consumer assistant formerly called Le Chat, so do not expect to find Large 4 in the Vibe model menu yet. Mistral says ML4 will run in several regions, including a European deployment it operates end to end under European law. Private cloud and on-premise use follow once the weights are out. If you mainly want to compare answers from several frontier models without managing API keys, Fello AI puts Mistral next to Claude, GPT and Gemini in one Mac app.

Should You Switch to Mistral Large 4?

Switch now only if European hosting or cyber work is the reason you are shopping. For a team that needs its AI processed in the EU under European law, Mistral Large 4 is the strongest model Mistral has ever served from its own European infrastructure. No other European model comes close on Artificial Analysis. For security teams tired of refusals on legitimate vulnerability work, Mistral's reproduce-and-patch result is the clearest reason to test it this month.

For everyone else, wait for October 27.

On today's independent numbers, GLM-5.3 and Kimi K3 are still the stronger open models, and you can download both now. The preview is a moving target by Mistral's own account, the licence is unknown and the half price may not last. The real verdict on Le Chonk comes when the weights land and outsiders can test the final checkpoint on their own hardware.

FAQ