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.

OpenRouter· Models· 1 discussions

OpenRouter adds Crouton Alpha, a free stealth model with a 1M-token context

OpenRouter announced a new stealth model, Crouton Alpha, which it describes as "a highly efficient reasoning model built to scale out the work your agents repeat all day". The maker is not named.
Coding

Apple opens Siri AI to all developers in the 27.3 betas

Apple opened a developer preview of Siri AI to all developers with the iOS, iPadOS, macOS and visionOS 27.3 betas, according to Apple Developer news as relayed by Techmeme. Any app can bring its content and actions into Siri AI through App Intents, including multi-step skills.
OpenAI· Coding

Unity says its official plugin for OpenAI's Codex is in OpenAI's plugin directory

Unity's games account said it launched an official Unity plugin for Codex, available in OpenAI's universal plugin directory. According to Unity, once the plugin is installed Codex can choose any skill in it to help complete a task correctly, and Unity published a walkthrough video of four examples. OpenAI's developer account reposted the post.
Coding

ai-cli gives any coding agent image, video and audio generation

Chris Tate of Vercel introduced ai-cli (npm i -g ai-cli), a command-line tool that lets an agent generate images, video and audio with any model. His post lists it working with Claude Code, Codex, OpenCode and Vercel's fx harness, with models such as Opus, Flux, GPT, Gemini, DeepSeek, GLM and Grok.
Anthropic· Products

Claude can auto-connect plugins when users log in, in beta on Team and Enterprise

Anthropic's ClaudeDevs account says users can now have plugins connect automatically when they log in to Claude, if the plugin's MCP server accepts Claude as a trusted issuer. Anthropic says this lets users get started faster and stay connected without re-authorizing, and asks plugin builders to add support.
Google· Models

Gemini 2.5 Pro and 2.5 Flash are to be deprecated on December 1 and redirected to newer models

Patrick Löber, who posts about the Gemini API, says Gemini 2.5 Pro and Gemini 2.5 Flash will be deprecated on December 1 in the Gemini API, and requests to them will be redirected automatically:
Coding· 1 discussions

Users report Postman's desktop app silently installs Codex and Claude Code plugins; not confirmed by Postman

A GitHub issue opened on October 25 on Postman's support tracker says Postman desktop 12.33.4 installed plugins into Codex and Claude Code on its own, after the user opened the app on a Mac. The reporter's evidence is Postman's own log, which records an "auto-existing-upgrade" run with "claude-code:installed" and "codex:installed" on October 24. The issue was widely shared on Hacker News on November 3.
Infrastructure

SurrealDB 4 launches as "the database and memory for every agent", with a serverless SurrealDB Earth on Cloudflare

SurrealDB announced SurrealDB 4 on X, calling it the database and memory for every agent. With it comes SurrealDB Earth, a serverless version of SurrealDB built on Cloudflare that, SurrealDB says, scales to as many agents and users as needed, wherever they are. SurrealDB says Earth can be started free with no card.
OpenRouter· Models

OpenRouter's catalog gives five MiniMax models an expiration date of November 5

OpenRouter's model catalog now shows an expiration date of November 5, 2026 for five MiniMax models: MiniMax-01 (minimax/minimax-01, 1,000,192-token context), MiniMax M2 (minimax/minimax-m2), M2.1 (minimax/minimax-m2.1) and M2.7 (minimax/minimax-m2.7), each with a 204,800-token context, and M2-her (minimax/minimax-m2-her, 65,536 tokens). None of the entries had an expiration date before. This is a catalog field only; the entries do not say why or what happens to the models after that date.
Infrastructure

Vercel starts rolling out Device Bound Session Credentials for vercel.com

Malte Ubl of Vercel says the company is starting to roll out Device Bound Session Credentials (DBSC) for vercel.com. He says DBSC makes it impossible to steal session cookies from a device, so info-stealer malware, for example installed through a supply chain attack, could not take a user's vercel.com session.
Google· Products

Schools that block Google's AI will lose student access to its search engine, the New York Times reports

The New York Times reports, in an article headlined "Schools that block Google's A.I. will lose student access to its search engine", that schools that block Google's AI features will lose their students' access to Google Search.
GitHub Copilot· Coding

GitHub Copilot is now available in JetBrains Air

GitHub says you can now delegate coding tasks to Copilot, guide their execution and review results inside JetBrains Air. The integration uses the Agent Client Protocol (ACP) and works with tasks in a JetBrains IDE or in the cloud, alongside the existing GitHub Copilot plugin.
OpenAI· Products

Users report OpenAI re-enabling model training on Codex chats; not confirmed by OpenAI

A Hacker News post titled "OpenAI seems to be silently re-enabling model training on Codex chats" says the setting that lets OpenAI use Codex conversations to train its models has turned itself back on for some accounts. The post links to no evidence.
Anthropic· Models

Anthropic's amended S-1 says it may soon meet its own definition of accelerated AI progress

An article on epyhia.com, which cites TechCrunch, reports that Anthropic's revised S-1/A registration statement, filed on Monday, November 2, includes a new risk disclosure. According to the article, the filing says Anthropic's internal estimates and expectations about future models, not its current ones, indicate it may "imminently" meet the definition of "dramatically accelerated progress attributable to AI" in its Responsible Scaling Policy, a threshold the article equates with recursive self-improvement. The filing says none of its current models has crossed that threshold and that Anthropic is considering changes to the policy "to keep humans in control of frontier model development at Anthropic."
OpenAI· Models

OpenAI says GPT-6 Astra can design a device's enclosure and circuit board, and seeks engineering partners

OpenAI describes how its models help with mechanical and electrical engineering: CAD models, manufacturing drawings, simulations and PCB design. These are OpenAI's own claims.
OpenRouter· Models

OpenRouter's catalog removes the November 3 expiration date from Meta: Llama 3.1 70B Instruct

OpenRouter's model catalog no longer shows an expiration date for "Meta: Llama 3.1 70B Instruct" (meta-llama/llama-3.1-70b-instruct) on November 3, 2026, the day the model was due to expire. The field had read November 3 since late October and is now empty. The entry still lists a 131,072-token context. OpenRouter has not said whether the retirement is cancelled or only postponed.
Anthropic· Coding

Claude Code 2.1.317 asks before sending your email address to the network and adds a gateway cost report

Claude Code v2.1.317, released November 2, 2026, per its release notes:
Moonshot AI· Models

Artificial Analysis: Kimi K3.1 Pro's personality differs more from Kimi K3 than any model update it has measured

Artificial Analysis says Kimi K3.1 Pro shows the biggest personality change it has seen between a model and its direct predecessor on AA-PersonalityBench. Each model is percentile-ranked from 1 to 100 on 30 traits. The scores are the site's own.
Coding

Vercel MCP shows deployments as an interactive card in ChatGPT

Vercel says its MCP server now shows deployments as an interactive card in ChatGPT and in other clients that support MCP Apps. The card shows live deployment status and build logs that can be debugged in the chat, and a Visit button appears once the URL is ready. Vercel describes the change in its changelog.
Models

Perplexity releases pplx-embed-v2-context, open-weight vision-language embedding models in 9B and 0.6B

Perplexity released pplx-embed-v2-context, two open-weight vision-language embedding models (9B and 0.6B) on Hugging Face, with a paper written with turbopuffer. Both return one int8 vector per document, or one per chunk that sees the whole document, and both also run in a non-contextual mode with one embedding per document and no chunking.