AI TrackerGet the app

AI news,
sorted for builders.

Launches, price changes and retirements from every AI provider, summarized. The same feed as the app's News tab.

Models

Axiom says AI agents formalized the Classification of Finite Simple Groups in Lean

Axiom, an AI math company, announced on X a complete Lean auto-formalization of the Classification of Finite Simple Groups, a result it calls the "periodic table of symmetry". It describes the original proof as about 15,000 pages written by over 100 mathematicians across generations, and calls it the largest math proof in history to be computer-verified.Models

Ling releases the open weights of Ling-3.1-flash-1019 and a base model

The Ling team says Ling-3.1-flash-1019 is now open source, with weights on Hugging Face and ModelScope. It also released Ling-3.1-flash-base, which it describes as an open foundation for further training. The team says the new model had further post-training for reasoning and real-world agent workflows.OpenRouter· Models

OpenRouter's catalog gives Meta: Llama 3.1 70B Instruct an expiration date of November 3

OpenRouter's model catalog now shows an expiration date of November 3, 2026 for "Meta: Llama 3.1 70B Instruct" (meta-llama/llama-3.1-70b-instruct), a model listed since July 23, 2024 with a 131,072-token context. The entry had no expiration date before. This is a catalog field; neither OpenRouter nor Meta has given a reason in what we have seen, and the entry does not say what happens to the model after that date.OpenCode Go· Coding

OpenCode adds plugins for OpenCode Desktop

LukeParkerDev announced on X on October 27 plugins for OpenCode Desktop, with a short demo video and a link to the OpenCode documentation on building GUI plugins. thdxr, who quoted the post, said the desktop plugins extend the same pattern as OpenCode's TUI plugins, so one package can ship server, TUI and desktop plugins, and called the right-hand panel a good place for useful additions. The post was also reshared by the OpenCode account.Models

Maya-2, a voice model for Indian languages, is released with open weights

Maya-2, a voice (speech synthesis) model aimed at India, was published with open weights on October 27 under the Hugging Face name maya-research/maya2-native. The announcement on X says it covers more than 1.5 billion people across all 28 states and 8 union territories.OpenAI· Coding

OpenAI and Oracle expand their partnership to bring Oracle apps into ChatGPT

OpenAI says it is expanding its partnership with Oracle to bring Oracle applications and data into ChatGPT and to power new Oracle products with OpenAI models. According to OpenAI, Oracle is developing Oracle AI Workspace, which would let people find information and complete approved business workflows across Oracle finance, procurement and planning applications from ChatGPT, within each organization's permissions and approvals.Anthropic· Models

A paper deposited by Anthropic claims an internal Claude model also refuted the Exponential Time Hypothesis

A preprint on hexagonmath.org, listed as deposited by Anthropic on October 27, 2026, claims a deterministic algorithm for 3-SAT on n variables that runs in 2^o(n) time, which would refute the Exponential Time Hypothesis (ETH). It says the algorithm also solves the harder #3-SAT problem, which counts satisfying assignments, and that the result has been formalized in Lean 4.Models

Baidu open-sources Unlimited OCR 2, a document parser with near-constant decoding cost

Baidu announced Unlimited OCR 2 on X on October 26, with code on GitHub and weights on Hugging Face. The model is built to parse documents of virtually any length within a finite context window.GitHub Copilot· Coding

GitHub Copilot CLI 1.0.104 makes Autopilot wait for background tasks and opens Assisted Permissions

GitHub Copilot CLI 1.0.104, dated October 27, 2026, per its release notes:OpenRouter· Models

OpenRouter's catalog drops the December 31 expiration date for Z.ai: GLM 4.5V

OpenRouter's model catalog no longer shows an expiration date for "Z.ai: GLM 4.5V" (z-ai/glm-4.5v). The entry had listed December 31, 2026 since October 24. The model, listed since August 11, 2025 with a 65,536-token context, remains in the catalog. This is a catalog field; neither OpenRouter nor Z.ai has given a reason for adding or removing the date.Anthropic· Infrastructure

Anthropic's Logan Graham corrects the Washington Post account of the Cathedral intrusion

Logan Graham of Anthropic's Frontier Red Team posted on October 26 corrections to the Washington Post story and the reaction to it on X, which said a security startup's AI agents got into Anthropic's systems in under 12 hours. These are his statements, not an Anthropic announcement:Z.ai· Infrastructure· 2 discussions

Cloudflare says a GLM 5.3 setup mismatch on Workers AI exposed KV cache pages across requests

Cloudflare published a disclosure and remediation write-up on October 26 saying a configuration mismatch in GLM 5.3 (@cf/zai-org/glm-5.3) on Workers AI triggered a bug in SGLang, the inference engine, that exposed KV cache pages across requests. The disclosure was coordinated with a post from SGLang.GitHub Copilot· Coding

GitHub Copilot assisted permissions are generally available in the CLI, the Copilot app and VS Code

GitHub says assisted permissions are now generally available in GitHub Copilot CLI, the GitHub Copilot app and Visual Studio Code. When Copilot asks to run a command, access a file system path or open a URL, a large language model evaluates the request. Requests it classifies as clearly safe proceed automatically; all others need the user's approval.Models

Artificial Analysis launches Code2Video-AA, a benchmark of marketing videos made in code

Artificial Analysis introduced Code2Video-AA, its own implementation of HeyGen's Code2Video Bench, built in partnership with HeyGen. Models get 114 product and brand marketing briefs and must answer each with a HyperFrames HTML composition in a single response, without tools; the composition is rendered to video with no network access. A judge model (Gemini 4 Argon, high) compares pairs of videos on engagement, prompt intent, composition, temporal consistency and craft, and the five axis Elo ratings are averaged into the headline score. More than 35,000 verdicts rank 27 models. The results are Artificial Analysis's own.Cursor· Models· 1 discussions

DeepSeek V4.1 Flash is now available in Cursor

Cursor announced on October 26 that DeepSeek V4.1 Flash is available in the editor; it can be turned on under Cursor Settings > Models. Cursor says every effort level sits on the CursorBench cost frontier, at $0.60 to $0.98 per task, with no model scoring higher at that cost. These are Cursor's own figures from its own benchmark.Anthropic· Coding· 2 discussions

Claude Code 2.1.311 changes ultrathink, loads model settings from a signed file on cloud providers, and fixes the 1M-context error

Claude Code v2.1.311, released October 26, 2026, per its release notes:OpenAI· Models· 7 discussions

OpenAI says its models proved the local smoothing conjecture, refuted the Exponential Time Hypothesis and proved Yang–Mills existence

OpenAI announced three results in mathematics, computer science and physics produced by its models. Everything below is OpenAI's own claim; none has been independently verified yet, and OpenAI says the results were checked in Lean.Coding

Scale AI's AgentEnv evals can now run on Vercel Sandbox

Vercel says you can now run AgentEnv evals from Scale AI Labs on Vercel Sandbox. You give an agent a task, run it in a sandbox and check how it performed. Vercel published a changelog entry and a setup guide in its sandbox documentation.Coding

Vercel Sandbox commands now support stdin

Vercel says commands run in Vercel Sandbox can now receive input on stdin. That lets you send input to running CLI tools and to stdio MCP servers, such as agent-browser, inside a sandbox. The details are in Vercel's changelog entry.GitHub Copilot· Models

GPT-6.1 Terra is now generally available in GitHub Copilot

GitHub says GPT-6.1 Terra, OpenAI's newest model, is generally available in GitHub Copilot. GitHub describes it as built for everyday agentic coding and terminal work at a lower cost than GPT-6.1 Sol, and says that in early testing it solved noticeably more terminal and C# coding tasks than GPT-5.6 Terra, working in focused, consolidated steps on well-scoped tasks.
Older news