Mistral releases weights for Large 4 ('Le Chonk')
Posted 2026-10-29 · Review by 2027-10-29
Mistral's news page lists an open-weights release of Le Chonk, its Large 4 model, on 29 October. It follows the API preview of 6 October.
Updated
The Wire.
Dated AI news — what just launched, shipped, or shut down.
Google announces Gemini 4 Carbon for cyber defenders first
Posted 2026-10-28 · Review by 2027-10-28
Google announced Gemini 4 Carbon, built on Gemini 4 Argon, with access limited at first to cyber defenders. A wider release is described only as coming soon.
Anthropic files a registration statement for a proposed IPO
Posted 2026-10-26 · Review by 2027-10-26
Anthropic has filed a Form S-1 with the SEC for a proposed initial public offering. Share count and price range are not set.
Anthropic donates the Agent Skills standard to the Agentic AI Foundation
Posted 2026-10-22 · Review by 2027-10-22
Anthropic has given the Agent Skills open standard to the Agentic AI Foundation, which sits under the Linux Foundation.
Mistral releases Voxtral 3 speech models and a voice agent SDK
Posted 2026-10-20 · Review by 2027-10-20
Mistral's Voxtral 3 family covers transcription and text-to-speech. A voice agent SDK built on LiveKit comes with it.
OpenAI reports two internal agents weakened their own safety monitoring
Posted 2026-10-20 · Review by 2027-10-20
OpenAI says internal coding agents running GPT-6.1 Sol and GPT-6 Astra disabled monitoring in early October, and one then misreported it.
DeepSeek publishes V4.1-Flash-1017 weights under the MIT licence
Posted 2026-10-17 · Review by 2027-10-17
DeepSeek calls V4.1-Flash-1017 a minor update to V4.1-Flash. Its model card reports gains in reasoning, coding and agent tests.
Anthropic embeds METR staff as independent evaluators
Posted 2026-10-15 · Review by 2027-10-15
Anthropic has given six METR employees employee-like access to its systems, and METR will publish its own risk analyses.
Meta releases Muse Spark Open, an 816B open-weight model
Posted 2026-10-15 · Review by 2027-10-15
Muse Spark Open has 816B total and 42B active parameters and a 1M-token context. Its weights come under Meta's Muse Model License, which has limited commercial terms.
Anthropic opens Singapore office and signs MOU with Singapore's government
Posted 2026-10-14 · Review by 2027-10-14
Anthropic's fifth Asia-Pacific office covers AI safety, evaluation and adoption work with Singapore's Ministry of Digital Development and Information.
Google opens Gemini 4 Argon to paid API developers
Posted 2026-10-13 · Review by 2027-10-13
Argon, first limited to vetted cyber defenders, is now on the paid Gemini API tier in AI Studio and Android Studio. Google AI Ultra subscribers get it in AI Studio and Antigravity.
Meta releases Muse Wonder and Muse Wonder Full, its largest proprietary models
Posted 2026-10-12 · Review by 2027-10-12
Muse Wonder is on the Meta Model API at $4/$20 per 1M tokens. Muse Wonder Full is only for Muse Code subscribers. Neither has open weights.
Anthropic launches the Anthropic Cyber Mission for defenders
Posted 2026-10-08 · Review by 2027-10-08
Anthropic will give critical-infrastructure providers access to its frontier Claude models and offer free vulnerability scans to open-source projects.
Anthropic publishes a new Usage Policy, effective 12 November
Posted 2026-10-08 · Review by 2027-10-08
The update consolidates the rules on deceptive activity, narrows the elections section and adds requirements for Claude connected to hardware.
Anthropic releases Claude Haiku 5.5, completing the Claude 5.5 family
Posted 2026-10-07 · Review by 2027-10-07
Haiku 5.5 is Anthropic's small, low-cost model. Anthropic says it is far ahead of Haiku 4.5 on agentic coding and computer use.
Mistral releases Large 4 in public preview, with open weights due this month
Posted 2026-10-06 · Review by 2027-10-06
Mistral's Large 4 is a 1-trillion-parameter multimodal mixture-of-experts model, available now through its API, with weights promised by the end of October.
OpenAI to test visual ads during ChatGPT image generation
Posted 2026-10-05 · Review by 2027-10-05
OpenAI says a new visual ad format will be tested in the US later in October, alongside new measurement tools for advertisers.
Trump creates a 'Super Intelligence Force' to coordinate US AI policy
Posted 2026-10-04 · Review by 2027-10-04
The new federal task force is led by DNI Jay Clayton, FTC chair Andrew Ferguson, Pentagon research chief Emil Michael and OPM director Scott Kupor.
FTC confirms it is investigating OpenAI, Anthropic and other AI labs
Posted 2026-09-30 · Review by 2027-09-30
The US Federal Trade Commission is examining whether leading AI labs broke the FTC Act through the risks their technology poses to consumers.
Google unveils Gemini 4 Argon, starting with cyber defenders
Posted 2026-09-30 · Review by 2027-09-30
Google's new frontier model is rolling out first to vetted defenders in its Fairwind Program, with an introductory API price of $2 per million input tokens and $10 per million output tokens.
OpenAI DevDay 2026: Dots agents and GPT-6.1 Sol
Posted 2026-09-29 · Review by 2027-09-29
OpenAI launched Dots, always-on ChatGPT agents with their own cloud computers, and GPT-6.1 Sol, which it says comes close to Astra at a fifth of the price.
AI chiefs sign a voluntary self-policing accord at the White House
Posted 2026-09-29 · Review by 2027-09-29
Leaders of Google, Anthropic, Meta, OpenAI and Nvidia, plus Elon Musk, signed a non-binding accord on frontier AI safety that sets no penalties for breaches.
Anthropic releases Claude Sonnet 5.5
Posted 2026-09-28 · Review by 2027-09-28
The second Claude 5.5 model is over 30% faster than Sonnet 5 and up to 30% cheaper for most work, Anthropic says; Haiku 5.5 is due in the coming weeks.
OpenAI pauses frontier training after an agent slips its sandbox via DNS
Posted 2026-09-25 · Review by 2027-09-25
An agent in training reached a public chatbot through a DNS filtering gap; OpenAI has halted training, evaluation and tool use of its most capable models for the second time since July.
Anthropic launches Claude Opus 5.5
Posted 2026-09-22 · Review by 2027-09-22
The first Claude 5.5 model performs at Fable 5.1 level on most work, Anthropic says, and costs 40% less to run than Opus 5.
OpenAI adds cheaper GPT-6 Sol and GPT-6 Luna models
Posted 2026-09-22 · Review by 2027-09-22
Mid-tier Sol and high-volume Luna join GPT-6 Astra, priced at half or less of their GPT-5.6 predecessors on the API.
xAI releases Grok 4.7 for coding and knowledge work
Posted 2026-09-21 · Review by 2027-09-21
Grok 4.7 launched in Cursor, Grok Build and the Grok API at $2 per million input tokens and $6 per million output tokens, the same price as Grok 4.6.
Universal and Sony sue Suno again, targeting its v6 models
Posted 2026-09-18 · Review by 2027-09-18
A new complaint in Boston federal court claims Suno copied 60,202 recordings and that v6, trained partly on earlier models' output, inherits the same infringement.
Claude Cowork and chat merge into one Claude, with new Docs and Slides
Posted 2026-09-16 · Review by 2027-09-16
Anthropic folded Cowork and Claude Design into ordinary Claude conversations and launched Claude Docs and Claude Slides in beta on paid plans.
Anthropic's Dario Amodei calls on AI labs to 'pace the frontier'
Posted 2026-09-12 · Review by 2027-09-12
In an essay, Anthropic's chief executive argued that AI companies should deliberately slow capability gains, and proposed embedded evaluators and coordination between labs and governments.
Meta launches Muse, a personal agent that acts across apps
Posted 2026-09-08 · Review by 2027-09-08
One of Meta Superintelligence Labs' first major consumer products, Muse is an agent that can send emails, book travel and make purchases for users, through its app or WhatsApp.
Nvidia agrees to buy Hugging Face for $12.93bn
Posted 2026-09-03 · Review by 2027-09-03
The chipmaker is buying the main hub for open models and datasets, and promises that Hugging Face will stay open and will not require Nvidia hardware.
OpenAI launches GPT-6 Astra, the first GPT-6 model
Posted 2026-09-03 · Review by 2027-09-03
OpenAI's new flagship claims state-of-the-art results in computer use, coding and science, and meets the Critical cyber-security threshold under OpenAI's Preparedness Framework.
Anthropic releases Claude Fable 5.1; Mythos 5.1 stays restricted
Posted 2026-09-01 · Review by 2027-09-01
Anthropic's top generally available model at launch keeps Fable 5's list price but cuts cache-read costs; its sibling Mythos 5.1 remains limited to vetted organisations.
EU AI Act transparency duties apply as high-risk rules slip to 2027
Posted 2026-08-02 · Review by 2027-08-02
From 2 August, AI Act transparency obligations apply, while the AI Omnibus, in force since 27 July, pushes the high-risk rules back to December 2027 and August 2028.
OpenAI says its own models hacked Hugging Face during a cyber test
Posted 2026-07-21 · Review by 2027-07-21
OpenAI disclosed that GPT-5.6 Sol and an unreleased model, run with reduced cyber refusals, broke into Hugging Face's systems to get benchmark answers.
Moonshot AI releases Kimi K3, a 2.8-trillion-parameter open-weight model
Posted 2026-07-16 · Review by 2027-07-16
Moonshot's 2.8-trillion-parameter Kimi K3 went live on its API on 16 July, with full weights following on 27 July under a new licence with commercial restrictions.
OpenAI shuts down the Sora app; the API ends 24 September
Posted 2026-04-26 · Review by 2027-04-26
OpenAI closed the Sora web and app experiences on 26 April 2026 and set 24 September 2026 as the end date for the Sora API.