AI news: week of Nov 2–8, 2026

Headlines

• 5 NovMeta Meta adds voice conversations and expressive avatars to the Muse appMuse app users in the US can now talk to their Muse by voice and see an animated avatar that reacts during the conversation. The rollout covers mobile, web and Mac, and voice on AI Glasses and Muse Charm is coming soon.
• 5 NovGoogle Google releases Gemini Omni 1.2 Flash with 15-second video and a 30% lower priceGemini Omni 1.2 Flash is generally available in the Gemini API and Gemini for business, with native video generation and editing up to 15 seconds (up from 10), multi-turn conversational editing and better reference following. The price is 30% lower than Omni 1.1, at $0.07 per second for 720p video.
• 5 NovMistral Mistral expands Studio with managed agents, apps and new workflow toolsMistral's enterprise platform Studio adds managed agents and apps plus new features for workflows, enterprise context, agent observability and evaluations. A new Mistral MCP server lets developers build locally and deploy to Studio with one command.
• 5 NovOpenAI OpenAI publishes community compact and policy blueprint for AI compute campusesOpenAI introduced a National Community Compact and a Compute Campus Policy Blueprint, which it says are meant to protect local resources, create jobs and guide how AI infrastructure is built.
• 4 NovOpenAI OpenAI to acquire Pierre Computer Company to expand Codex code toolsOpenAI plans to acquire Pierre Computer Company, Inc., a deal it says will expand the Codex tools for building, understanding and reviewing code.
• 4 NovOpenAI OpenAI pipeline finds 948 Linux kernel privilege-escalation bugs and drafts patchesOpenAI built an autonomous pipeline that found and validated 948 local privilege-escalation vulnerabilities in the Linux kernel and generated patches for them.
• 4 NovOpenAI Daybreak Red with GPT-5.6 Cyber comes to AWS GovCloud (US-East) via Amazon BedrockDaybreak Red is now available in AWS GovCloud (US-East) through Amazon Bedrock, giving approved customers access to GPT-5.6 Cyber for authorized cybersecurity work.
• 4 NovOpenAI OpenAI releases Images 2.6 with faster generation, GIFs and more precise editingImages 2.6 brings faster image generation, GIF creation, more precise editing and new API controls. It is available in ChatGPT, Codex and the API.
• 3 NovOpenAI OpenAI names Jill Hazelbaker Chief Corporate Affairs OfficerJill Hazelbaker is joining OpenAI as Chief Corporate Affairs Officer, a role that leads the company's global communications and corporate positioning.
• 2 NovAnthropic Claude API adds message threads in beta, so apps send only new messages each turnThe Claude Messages API can now keep a conversation on the server between turns, so an app sends only the new messages and the previous message ID, and Claude starts responding sooner. The feature is in public beta.
• 2 NovOpenAI OpenAI raises Security Bug Bounty rewards and adds researcher bonusesOpenAI is raising the rewards in its Security Bug Bounty program and introducing bonuses for report quality, loyalty and priority campaigns for independent security researchers.
• 2 NovOpenAI OpenAI introduces LifeVal, a benchmark for AI help with everyday tasksLifeVal is a new OpenAI benchmark that measures how well AI helps people complete everyday tasks, such as finding childcare, managing finances and handling government paperwork.
• 2 NovAnthropic Anthropic releases FlashProve, an open-source prototype for provable LLM inferenceAnthropic's Frontier Red Team published FlashProve, a research system that uses zero-knowledge proofs so users can check that a response came from the model a provider committed to, without seeing its weights. Code, Lean proofs and a technical report are out under Apache 2.0.
• 2 NovOpenAI OpenAI names Dan Driscoll Senior Counselor for National Security ModernizationOpenAI has appointed Dan Driscoll to advise on responsible AI for national security modernization, cyber resilience and collaboration with U.S. allies.
• 2 NovxAI xAI retires grok-imagine-image-quality and serves it with a cheaper Imagine 2.0 tierRequests to grok-imagine-image-quality and grok-imagine-image-pro now go to grok-imagine-image-2.0 with quality set to low, at a lower per-image price. xAI says no request changes are needed.

Announcements on X

Our one-line summaries of selected posts, each linked to the original. Not the full account timelines.

• 5 NovCodex Tibo says OpenAI's latest cyber-specialized model is out, available only through Daybreak Red access; the post does not name it.@thsottiaux · View on X
• 5 NovCodex Tibo says dots have just received a large set of improvements; the post does not list them.@thsottiaux · View on X
• 4 NovClaude Background subagents in Claude Code use less of the usage limit: Haiku 5.5 now writes their 30-second status updates from short notes instead of re-reading the whole conversation, about 99% cheaper per update.@ClaudeDevs · View on X
• 3 NovClaude Claude can now connect plugins automatically when users log in, for MCP servers that accept Claude as a trusted issuer; plugin builders are asked to add support.@ClaudeDevs · View on X
• 2 NovClaude Message threads enter public beta on the Claude Messages API: the conversation stays on the server between turns, so apps send only new messages and the previous message ID.@ClaudeDevs · View on X
• 2 NovCodex Tibo says OpenAI is adding more frontier models to dots, including GPT-6.1 Terra alongside GPT-6 Astra, to make dots faster and reach more people.@thsottiaux · View on X

Claude Code releases (2)

• 4 NovClaude Code 2.1.318Added a record of your answer when Claude Code asks to turn on the mods it writes in a session · Added assume_role on…
• 3 NovClaude Code 2.1.317Added a warning to /schedule when a new cloud scheduled task gets connectors you did not ask for, with how to remove them ·…

Claude Code changelog