Mistral Large 4 (Le Chonk) arrives on Hugging Face: a 1T-parameter multimodal MoE open-weight model
Hugging Face published a launch blog for Mistral Large 4 (Le Chonk), a multimodal, multilingual MoE model with about 1T total parameters, 49B active and a 1M token context window.
模型发布/更新
-
Mistral Large 4 (Le Chonk) lands on Hugging Face: 1T-parameter MoE multimodal open-weight model
Hugging Face published a launch blog for Mistral Large 4 (Le Chonk), a multimodal multilingual MoE model with about 1T…
-
Pika post-trains and open-sources Pika H3 Flex video generation model based on MiniMax H3
Pika Labs post-trained on MiniMax H3 to launch and open-source Pika H3 Flex, cutting sampling steps from 50 to 6,…
产品发布/更新
-
Codex doubles usage limits for all Pro and Business Premium tiers starting today
Tibo announced that starting today, all usage limits are doubled for the Pro 100/200/500 tiers and Business Premium.…
-
Factory releases Droid 1.0, an enterprise agent harness built on programmatic tool calling
Factory released Droid 1.0, a model-agnostic agent harness centered on programmatic tool calling (PTC), in which…
-
OpenRouter rebuilds default routing algorithm, cutting TTFT by 16.9% and 429 errors by 67%
OpenRouter announces a rebuilt default provider routing algorithm, moving from price-only sorting to a Routing Cost…
行业动态
-
纽约时报称中国情报人员瞄准英伟达 CEO 黄仁勋
据《纽约时报》报道,美国情报机构认为英伟达 CEO 黄仁勋身边一名高管为中国提供情报,此人是新加坡数据中心运营商 Megaspeed 的 Alice Huang(黄乐)。
-
Reddit 宣布转向以登录为主的体验,以应对 AI 搜索带来的流量下滑
Reddit 在 Q3 财报中宣布将转向以登录为主的体验,以保护数据免受 AI 抓取并提升广告变现。10 月搜索流量平均每周减少超 100 万次引荐,Q3 全球日活为 1.303 亿,其中美国日活降至 5070 万,同比降 2%。Q3 收…
-
Google Q3 search growth misses expectations as investors worry AI search monetization weighs on core business
Google's Q3 search revenue grew about 12%, below analysts' 15% expectation, and its market value fell by as much as…
-
404 Media 起底 Proton 内部 AI 转型之争:CEO Andy Yen 强推 AI 引发恐惧文化
404 Media 采访九名现任与离职 Proton 员工并查阅内部文件,报道 CEO Andy Yen 强推 AI 转型引发的内部冲突。10 月公司 90% 代码由 AI 编写,多使用 Anthropic 的 Claude;Yen 称今…
论文研究
-
OpenAI 详解如何检测前沿模型失准行为
OpenAI 发布长文,详解其针对前沿模型失准行为的多层监测系统,包括激活探针、基于模型的审查、调查智能体和自动化调查四层架构,并对 GPT‑6 Astra 及以上模型的全部 RL 运行、评估和内部 Codex 流量进行监测。
-
Trakkr study traces provenance of 1,285 AI-cited statistics
Trakkr Research asked ChatGPT, Perplexity, Gemini, Claude, and Google AI Overviews 200 questions requiring numbers and…
技巧与观点
-
OpenRouter tutorial: extracting tables from PDFs with LLMs
OpenRouter released a tutorial on how to send a PDF as a file along with a strict JSON schema to a model that can read…