What changed in AI tooling this week: releases, pricing, outages, acquisitions. Every item links to its source.
Starting in Copilot CLI 1.0.94 0 (announced Oct 7), /model lists tool calling, streaming capable models from a running Ollama instance next to the cloud ones, and you can add one for the current sess…
Released Oct 7 on Hugging Face alongside the experimental d1 omni 600M (text plus image or audio), d1 3B returns a typed answer in one forward pass instead of generating text, scoring 48.57 on the De…
Announced Oct 7, npx skills add google/skills skill retrieving developer knowledge teaches Claude Code, Cursor, Copilot, or Antigravity to pull fresh Markdown docs for Google Cloud, Firebase, Android…
Released Oct 7 for subagents and high volume work, it scores 72.4% on OSWorld 2.1 (offline subset) and 39.2% on Terminal Bench 4.0 vs 16.4% for GPT 6 Luna, and it's the first Haiku with adjustable ef…
Free during its stealth period for Pro and Enterprise teams with purchased AI Gateway credits, stealth/glyph cluster is a text only reasoning model for coding and long context work that you can pick…
Released Oct 7, anyone can now run shell commands in a sandbox with /sandbox or sandbox , MCP server config edits take effect between turns mid session, and enterprises get a permissions.limitTo sett…
Released Oct 7, the update switches the bundled and Amazon Bedrock catalogs to GPT 6.1 Sol, lets you sign in to MCP servers mid session with /mcp login , and brings multi agent V2 and Ultra reasoning…
Released Oct 7 in 0.6B and 9B sizes on Hugging Face, the Qwen3.5 based models keep one 128 dim vector per token for text, images, and rendered document pages, scoring 62.3% and 65.2% nDCG@10 on ViDoR…
JAZ, open sourced at jaz lang/jaz , gives the model a single invoke primitive in a Python REPL where its prompt and its own history are variables, so it can pass its full history by reference to a su…
Released Oct 6, the Qwen3.8 27B based model returns probabilities for yes/no, choice, and rubric score questions instead of text, lifting its Hugging Face Decision Index score from 56.4 to 61.56 and…
Apple and EPFL researchers found a single long coding agent session (Malena, built on OpenCode) matched or beat MLEvolve, AiScientist, Arbor, and ScienceFlow at every frontier backbone tested, medali…
Launched Oct 7 at playground.google for US users 18+, it lets you describe a game, tweak its rules, physics, and characters in chat, then keep it private, share a link, or publish it to a public gall…
Released Oct 6 in openai/math , the papers form 372 result families drawn from about 4,000 problems, averaging roughly three hours of ChatGPT Pro thinking each, with Lean formalizations for many (not…
Released in beta on Oct 6, POST /v1/decisions runs only on gpt 6 luna and turns text or images into typed answers (a 0 to 1 predicate probability, a choice from your fixed options, or a score against…
Announced Oct 6 at Team '26 Europe as part of its new Agentic Multiplayer Protocol, Rovo Work is a Rovo Chat mode that proposes a plan you edit, then runs for hours across Jira, Confluence, and conne…
Announced Oct 7 at Team '26 Europe, you can pick one of these coding agents in a Jira ticket's assignee field (or @mention it in a comment) and it works from the ticket and opens a pull request, with…
Announced Oct 6 at Team '26 Europe, the hosted server (OAuth 2.1 + PKCE, respects each user's existing permissions) now reaches Bitbucket Cloud PRs, semantic Code Search, Confluence whiteboards and d…
In early access since Oct 6 for Premium and Ultimate on GitLab.com and Self Managed, it checks npm, pip, Poetry, Maven, Gradle, and Bundler installs against policies for malware flags, vulnerability…
Claude for Google Workspace, a public beta on all paid plans since Oct 6, installs once from the Google Workspace Marketplace and reads and edits whichever doc, sheet, or deck you have open (with an…
Released Oct 6 for all Wiz Code customers, the scanner runs on the Atlas harness (No. 1 on CyberGym at its July launch) to find intent dependent flaws like broken access control and missing ownership…