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.

OpenCode Go· Coding· 1 discussions

OpenCode says you can sign in with a Claude subscription again

The OpenCode account announced on X on October 30 that users can now use their Claude subscription in OpenCode, linking a demo video from Daniel Smolsky ("subscriptions in @opencode are back"). The post had about 861 points and 68 comments on X within the hour.
OpenAI· Infrastructure

OpenAI announces Private Intelligence, stronger data controls for businesses

OpenAI announced Private Intelligence, a set of data protections for businesses that use its most capable models on sensitive work. OpenAI says it is building the protections in close collaboration with some of its largest customers.
Coding

Cloudflare can send alerts straight to Slack, email or coding agents

Cloudflare says alerts about metric spikes, DDoS attacks and more can now be sent directly to Slack, email, or to Claude Code, Cursor and Devin. The setup is described in the Workers Observability documentation under issue automations. The change comes as a Cloudflare team rebuilds the company's notifications and alerting experience, starting with a new alerts interface in the dashboard under Observability → Alerts.
OpenAI· Models

OpenAI tightens ChatGPT's violence safeguards and limits weapons information for teens

OpenAI announced stronger violence-prevention measures in ChatGPT and a commitment to a public Community Safety Report. Everything below is OpenAI's account; it says the safeguards begin rolling out "in the next few weeks".
AlertCursor· Coding· 10 discussions

A trojanized Claude Code extension was served from Cursor's extension marketplace

Cursor says it identified a security incident involving a modified Claude Code VS Code extension distributed through its extension marketplace, affecting at most 5,000 users. Cursor says the malicious version was not developed by Anthropic and that no other extension is affected.
Models

Artificial Analysis launches a leaderboard for HLE-Diamond, the refined Humanity's Last Exam

Artificial Analysis launched its HLE-Diamond leaderboard, its own evaluation of the refined 1,000-question version of Humanity's Last Exam, as part of v5 of its Intelligence Index. CAIS and Scale AI released HLE-Diamond after a year of cleaning and refining the questions.
Google· Models

Cognition's FrontierCode 2.0 puts Gemini 4 Carbon at 40%, behind Claude Opus 5.5, GPT-6 Astra and Claude Sonnet 5.5

Cognition, the maker of Devin, evaluated Gemini 4 Carbon and Gemini 4 Argon on FrontierCode 2.0, its coding benchmark that measures a model's ability to go beyond the spec. These are Cognition's own measurements.
Anthropic· Coding

Claude Code 2.1.314 adds a setting to block writes outside the working directories and idle compaction

Claude Code v2.1.314, released October 29, 2026, per its release notes:
Infrastructure

Artificial Analysis launches a Search Index that ranks search APIs for agents

Artificial Analysis launched its Search Index, which measures the performance, cost and latency of search APIs across 1,700 tasks from three search benchmarks: BrowseComp, DeepSearchQA and AA-Omniscience. All rows use the same base model and its open-source agent harness, Stirrup. It says agents are expected to search the web far more than humans do by the end of the year.
Anthropic· Infrastructure

Claude API raises the maximum request size from 32 MB to 128 MB

Anthropic's ClaudeDevs account says the maximum request size for Messages and Token Counting on the Claude API went from 32 MB to 128 MB. It says image-heavy requests, such as long computer-use sessions and large PDFs, were hitting 32 MB with plenty of context window left, and that this gives about 4x more room.
Infrastructure

Vercel Firewall gets separate controls for AI crawlers and AI assistants

Vercel says its Firewall managed rules can now block AI crawlers and AI assistants independently. A site can block bots that collect content for AI training while still letting AI assistants fetch its pages in response to user requests. The details are in Vercel's changelog entry.
OpenCode Go· Coding

OpenCode Go is now available to whole teams

OpenCode announced that OpenCode Go can now be used by an entire team. Its page lists three plans: Go at $10 a month, Go Plus at $40 a month with higher limits, and Go Teams at $20 per member per month with limits for every member. OpenCode describes Go as giving "generous usage limits and reliable access to leading coding models".
OpenAI· Models

OpenAI describes its expanded system for monitoring frontier models for misalignment

OpenAI published a look at the monitoring it now runs on frontier models. All of the following is OpenAI's own account.
OpenAI· Coding

Codex 0.166.0 adds keyboard transcript selection, `codex doctor --network-only` and GPT-6.1 Terra in more catalogs

Codex 0.166.0, per its release notes:
Google· Coding

Google says Jules, its asynchronous coding agent, is winding down

Google's Jules account says Jules will be winding down. It says Jules began as an experiment in how asynchronous coding agents can help developers ship faster, and that Google is working on bringing what it learned into Antigravity, to combine autonomous remote background task execution with an agent-first development environment. The post gives no date for the shutdown.
Anthropic· Models

Anthropic says it trained Claude to flag possible misuse in its own conversations

Anthropic described a safeguard it calls self-assessment: right after Claude responds, it gives a preliminary safety assessment of the conversation, flagging potential misuse. Everything below is Anthropic's own account, from its post and a new Safeguards blog.
Models

ElevenLabs releases SFX v3, a sound-effects model that follows multi-part prompts

ElevenLabs released SFX v3, its latest sound-effects model. It says the model makes richer, more precise sound effects from a single text prompt, follows several instructions in one prompt more reliably (its example: a crowd cheering, a ball dropping, then a crowd booing), and has improved music and vocals.
Models

PokeBot launches RoboAPI, an API that gives robot arms natural-language instructions

PokeBot introduced RoboAPI on October 29, powered by Poke-Aurora, which it calls its world-action model. It says compatible robots can take natural-language instructions through a simple API, with no task-specific training.
Models

Pika launches H3 Flex, a fast video model it calls the top open-weight text-to-video model on DesignArena

Pika introduced H3 Flex, which it says debuts as the #1 open-weight model on DesignArena's text-to-video board (Pika's claim). It describes the model as built for professional iteration: exportable audio stems, resolutions up to 4K and faster-than-realtime generation, with a 15-second clip and soundtrack in as little as 7 seconds.
Anthropic· Coding

Claude Code adds CLAUDE_CODE_SUBAGENT_EFFORT to set a lower effort for subagents

Claude Code's @bcherny said the environment variable CLAUDE_CODE_SUBAGENT_EFFORT has landed. Running CLAUDE_CODE_SUBAGENT_EFFORT=low claude sets the effort level subagents use, and it can also be set in a custom agent's config. He was replying to a user who asked whether subagents could default to a lower effort instead of inheriting the main session's.