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· Infrastructure

OpenRouter rebuilds its provider router to weigh quality against actual cost

OpenRouter says it shipped the biggest improvement to its provider router since its founding. Before, it says, routing prioritized mostly by price; the new router rewards the best-quality inference for the actual cost of a request, so requests stay cheap but also fast and reliable. OpenRouter says providers that deliver more value automatically win more traffic. It describes the change in a blog post, "How we rebuilt provider routing".Google· Coding

Antigravity adds a plugins marketplace

Google's Antigravity account announced a new plugins marketplace in Antigravity, saying it makes it easier to search for and discover customizations. The post had 532 likes on X by October 29.OpenAI· Models· 1 discussions

OpenAI updates its Model Spec around five core values

OpenAI published an updated Model Spec, its public description of how it wants its models to behave. The revision reorganizes the guidance around five values: faithfulness to human intent, harm minimization, honesty and transparency, human agency, and care. Everything below is OpenAI's account.OpenAI· Products· 2 discussions

OpenAI doubles usage limits on ChatGPT Pro 100, 200 and 500 and Business Premium

OpenAI's @thsottiaux, in a post that OpenAI's account reposted, said OpenAI is doubling all usage limits for the Pro 100, 200 and 500 tiers and for Business Premium, effective October 29. The post had about 9,000 likes and 2,200 replies on X by the same day.Products

Vercel Agent can investigate alerts before paging you

Vercel says Vercel Agent can now validate alerts before paging. It checks logs and metrics to assess the impact, then escalates only the alerts it marks as critical. Vercel describes the change in its changelog.OpenAI· Models· 33 discussions

OpenAI releases GPT-6.1 Terra, a cheaper computer-use model at half the price of GPT-5.6 Terra

OpenAI released GPT-6.1 Terra on October 26, 2026, describing it as a fast, capable computer-use model: it understands what is on screen, takes action across apps and writes naturally. It is rolling out to all Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex, and developers can use it in the API as gpt-6.1-terra. OpenAI says usage allowances in ChatGPT Work and Codex go further with it than with GPT-5.6.Models

Phonic 1 debuts at #1 on Artificial Analysis's Speech to Speech Index

Artificial Analysis (a benchmark site, its own measurements) says Phonic 1 (High), the speech-to-speech model from Phonic, takes #1 of 22 on its Speech to Speech Index at 87.1, narrowly ahead of Grok Voice 3.0 (86.9). The model is available through the Phonic API at $0.07 per minute.Models

Fish Audio releases Drama 3, which Artificial Analysis ranks #1 among open-weight speech models

Fish Audio released Drama 3, with weights on Hugging Face. Artificial Analysis (a benchmark site, its own measurements) says it is the #1 open-weights model on both of its speech arenas and #1 of all models on its Pronunciation Robustness benchmark.Coding

Paseo agents can now run on Vercel Sandbox

Vercel says Paseo agents now run on Vercel Sandbox. Vercel says Paseo installs with one command, lets users control sessions from their phone, and works with Pi, fx, Codex, Claude Code and other agents. Vercel published a changelog entry for it.OpenCode Go· Coding

OpenCode adds a /goal command

Neriousy announced on X on October 29 that the /goal command has landed in OpenCode, with a short demo video. The post does not say how the command works or how to turn it on. It was reshared by the OpenCode account and by thdxr; it had about 92 points and 12 comments on X shortly after it was posted.Moonshot AI· Products

Kimi adds Gmail, Google Drive and Google Calendar plugins

Moonshot's Kimi account says Kimi now works with Gmail, Google Drive and Google Calendar. According to Moonshot, you can summarize Gmail threads, find files in Google Drive and plan your day with Google Calendar, with OAuth access that you control. The plugins are listed on kimi.ai/plugins.Mistral AI· Models· 7 discussions

Mistral releases the weights of Mistral Large 4, a 1-trillion-parameter model, at about half the preview price

Mistral published the weights of an improved Mistral Large 4 ("Le Chonk"), its flagship natively multimodal model of 1 trillion parameters, which entered public preview earlier this month. Weights are on Hugging Face as Mistral-Large-4-1T-A52B and an NVFP4 version. Mistral calls it European-trained and the best open-weight model for coding. A commenter says the license is a modified MIT, not an OSI-approved one.RumorQwen· Models· 4 discussions

Qwen 4 is not announced, but Qwen Code already lists qwen4 model ids

An X account reports that Qwen Code v0.26.3, released on October 28, includes a change (pull request #15762 in the official QwenLM/qwen-code repository) that treats every model whose id starts with qwen4 as having a 1M-token context by default. According to that account, the tests name qwen4-max, qwen4-plus, qwen4-coder-plus, the dated qwen4-max-2026-10-01 and qwen4.5-max. The account's post links the pull request and shows a screenshot.Infrastructure

Ling and NVIDIA speed up Ling-3.0-flash on a single DGX Spark by 42.5%

The Ling team says that, working with NVIDIA, it optimized inference for the open-weight Ling-3.0-flash (124B total parameters, 5.1B active per token) on one NVIDIA DGX Spark. NVIDIA published a technical blog post on the work. The figures are the team's own.Anthropic· Coding

Users measure Claude Code's background-agent status summaries at 16–20% of requests; not confirmed by Anthropic

Flamerix AI published a write-up, shared on Hacker News (323 points and 100 comments by October 29), saying Claude Code spends a large share of its usage on a feature that updates the status line of background agents. It says that about every 30 seconds, for each background agent, Claude Code sends the agent's whole transcript to the agent's own model (usually Opus) and asks for a 3–5 word description of its latest action. The result only updates the agent's row in the agents panel.Anthropic· Coding

Anthropic finds no regression behind Claude Code usage complaints, and fixes a few token-burn bugs

Boris Cherny of the Claude Code team posted on X on October 28 that the team went through all the session IDs users shared after complaints that Claude Code was using up limits faster. He says they found a few smaller issues that made some sessions burn through tokens faster than they should, and that fixes are rolling out now.AlertCoding

Command Code revokes all API keys after a security incident tied to a distillation attack

Command Code, a coding agent, posted on X on October 28 that earlier that day it "detected and stopped a security incident linked to a large-scale distillation attack targeting Anthropic and OpenAI models." This is the company's own account.Infrastructure

Vercel introduces virtual models on AI Gateway

Vercel introduced virtual models on AI Gateway: a single virtual model ID that agents can share. Vercel says it lets teams manage model strings and configuration centrally, switch to new models immediately, and enforce Zero Data Retention.Models

Artificial Analysis shows AA-Robotics v0.1, a benchmark of models controlling a robot arm

Artificial Analysis published AA-Robotics v0.1, which has models control a robot arm and shows the task from the model's point of view. In an example where the same robot writes "hello world", Artificial Analysis says Claude Opus 5.5 dabbed it out dot by dot, with 3,270 arm moves over 34 minutes, while GPT-6 Astra wrote in continuous strokes, with 514 moves in 13 minutes.Models

Particle-1, a model post-trained on DeepSeek V4.1 Flash, is now on Vercel AI Gateway

Vercel says Particle-1 (particle/particle-1) is now available on AI Gateway, its unified endpoint for model providers. Vercel says the model is post-trained on DeepSeek V4.1 Flash and completes engineering tasks with fewer tokens, and suggests it for debugging, planning code changes and breaking down multi-step problems.
Older news