タグ: china

Qwen 3.8 is no longer a preview with a marketing claim attached to it. Alibaba made Qwen3.8-Max generally available on August 3, 2026, and this time the model arrived with the numbers it was missing in July: 95 billion active parameters per token out of 2.4 trillion total, a confirmed 1-million-token context window, and API pricing at $2 per million input tokens and $6 per million output tokens. That matters because the claim Alibaba led with, that Qwen 3.8 is “second only to Fable 5,” is finally testable. It turns out to be true on one leaderboard and wrong on another, and the gap between those two answers tells you […]

In January 2025, a single AI model from a little-known Chinese startup erased more than $1 trillion in market value from US tech stocks in one trading day. The model was DeepSeek-R1, and the man behind it is Liang Wenfeng, a former quant trader who built a frontier-grade AI lab on a fraction of the budget his American rivals spend. He rarely gives interviews, holds no flashy executive title beyond founder and CEO, and reportedly still reads research papers and writes code himself. So who is Liang Wenfeng, and how did a kid from a rural village in southern China end up reshaping the global AI race? This guide covers […]

Baidu released ERNIE 5.1 on May 8, 2026, hitting #4 on LMArena Search using only 6% of normal pre-training cost. We cover benchmarks, architecture, agentic performance, and how it compares to DeepSeek V4 Pro, Claude Opus 4.6, and Gemini 3.1 Pro.

MiniMax M2.5 scores 80.2% on SWE-Bench Verified, runs at 100 tokens/sec, and costs $1/hour. Here’s what we know — and what’s still unverified.

In September 2025, a Chinese state-sponsored hacking group reportedly used Claude Code (an AI programming assistant created by Anthropic) to carry out a large-scale espionage operation against roughly 30 global targets, including major tech companies, financial institutions, chemical manufacturers, and government agencies. Unlike previous AI-assisted attacks, the attackers built an automated framework around Claude Code so that the AI could handle 80–90% of the tactical work: scanning networks, finding vulnerabilities, writing exploit code, testing stolen credentials, and pulling down data. Humans mostly popped in at a few key checkpoints to approve big moves like exploitation and data exfiltration. This is believed to be the first documented case of an AI autonomously executing such […]

On November 6, 2025, Alibaba-backed Moonshot AI released Kimi K2 Thinking, its most advanced open-source model yet. It’s the first reasoning-focused variant in the Kimi K2 family and marks a major step forward in long-context, multi-step reasoning and autonomous tool use. Moonshot has moved fast since then. The Kimi line has advanced through K2.5, K2.6, and K2.7 to the 2.8-trillion-parameter Kimi K3, Moonshot’s newest flagship and its largest model yet, with the open weights expected to follow in late July 2026. Kimi K2 Thinking immediately made headlines for its performance: it set new state-of-the-art scores on several open benchmarks, including Humanity’s Last Exam (HLE) and BrowseComp, where it outperformed closed models like GPT-5 and Claude Sonnet 4.5. Unlike its competitors, K2 […]

Alibaba’s AI stack just crossed a symbolic frontier. On September 5 2025, the company’s cloud division unveiled Qwen-3-Max-Preview (Instruct), a 1-trillion-parameter large-language model that immediately takes its place among the most complex systems ever deployed and the most powerful AI models. The rollout began quietly inside Qwen Chat and Alibaba Cloud Model Studio. Soon coming to Fello AI. Only a few months ago, OpenAI’s GPT-4o, Google DeepMind’s Gemini 2.5, and Anthropic’s Claude Opus 4 each claimed slices of the “frontier-model” crown. Meanwhile, China-based challengers have been racing to close the capability gap. By releasing a one-trillion-parameter model, Alibaba is making it clear it wants to compete at the very top, not on the sidelines. […]

Moonshot AI, Chinese startup backed by Alibaba and Tencent, is rolling out an upgraded version of its open‑weight open-source large language model, Kimi K2. The new build Kimi K2‑0905, introduces a 256,000-token context window, improved coding performance, and retains its permissive modified MIT license. Although early beta access was briefly mentioned on the company’s Discord before details were removed, the model is now fully available to the public. Weights and a complete model card are live on Hugging Face, making it easy for developers to download, run, and fine-tune the model locally. Moonshot’s Meteoric Rise Founded in March 2023, Moonshot AI has grown at breakneck speed. In just over a year, the startup […]

The open-source AI race just got more interesting. Chinese startup DeepSeek has unveiled DeepSeek-V3.1, its biggest upgrade yet, bringing sharper reasoning, stronger coding skills, and new support for tool-calling and agent workflows. Unlike its earlier release that felt experimental, V3.1 arrives as a serious contender. With a 128K context window, hybrid reasoning modes, and an API that’s dramatically cheaper than its Western rivals, the model is designed to push open AI closer to the performance of GPT-5, Grok 4, Claude Opus, and Gemini 2.5 Pro. See our DeepSeek pricing guide for full API costs and free tier details. Early benchmarks show V3.1 not just catching up but competing head-to-head in coding and reasoning tasks, […]

中国の小さな新興企業が、ほぼすべての主要ベンチマークでグーグルの大注目製品Veo 3を凌ぐテキストからビデオへの変換モデルをリリースした。MiniMaxのHailuo 02です。物理シミュレーションからプロンプトの正確さまで、Hailuo 02はクリエイター、開発者、そしてインターネット全体を虜にしている。そしておそらく最も注目すべき点は？Veoの最上位プランが$250以上かかるのに比べ、Hailuo 02は月額わずか$8で誰でも利用できるのだ。このプラットフォームが広く知られるようになったのは、ユーザーが作成した、オリンピックスタイルのダイブをする猫の動画がTikTok、Reddit、Instagramで広まったことがきっかけだ。その動画は、中国のスタートアップMiniMaxの第2世代AIモデルであるHailuo 02を使って作られた。それ以来、このプラットフォームは37億以上の動画を生成している。TechRadarが報じたように、「Hailuo 02 [...]...