Model intelligence.
Follow the changes. Understand what comes next.
Google Launches Gemini 4 Neon in Antigravity With 1M Context, Four Thinking Levels, and Benchmarks vs. Claude Opus 5.5 and GPT-6.2 Sol
Announced on November 19, 2026 at 05:51 Beijing time, Google's Gemini 4 Neon features a 1M-token context window, four thinking levels, and planner-executor pairing with Gemini 4 Hydrogen in Google Antigravity, outperforming Claude Opus 5.5 and GPT-6.2 Sol across most of 18 published benchmarks.
Latest stories349
Newest publishedOpenAI adds Ultrafast mode to GPT-6.1 Sol: up to 8x faster, at six times the price
Ultrafast for GPT-6.1 Sol is open to all API users and rolling out in Codex and ChatGPT Work. OpenAI says it runs up to 8x faster than Standard; API prices are six times Standard, and in Codex only Pro $500 and eligible Enterprise and Edu plans get it.
Google Cloud launches Gemini agent, one work agent that picks between Gemini and Claude
Google Cloud's new Gemini agent works inside Gmail, Docs, Sheets, Microsoft 365 and Slack, sends each task to a Gemini or Claude model, and can run as a "coworker" with its own email address. It is in private preview for enterprise customers.
GPT-6 reaches ChatGPT Free and Go: Luna for free users, Sol for paid, Codex unchanged
OpenAI is rolling out GPT-6 with Intelligent UI to ChatGPT Free and Go users from October 8, a day after paid plans. Free and Go get GPT-6 Luna, paid plans get GPT-6 Sol, and the models behind ChatGPT Work and Codex stay the same.
Tibo says paid Codex users are getting another weekly usage reset
Tibo, who leads Codex and ChatGPT, posted on X that another reset is being sent to all paid accounts in the coming hours. In this context, reset means a weekly usage reset for paid Codex users, so people can use up their current quota before the new free reset lands.
Anthropic launches Claude Haiku 5.5, a cheaper and more capable small model
Anthropic has released Claude Haiku 5.5, the newest version of its smallest and lowest-cost Claude model. The company says it is faster and more capable than Haiku 4.5, while cutting average running costs by about 75%, with the biggest savings on requests under 100,000 tokens.
Tibo, Head of Codex & ChatGPT, completes Codex usage reset: prior updates earned praise, reset followed community vote
Tibo, Head of Codex and ChatGPT, has completed a Codex usage reset. It follows four widely well-received features and math proof capabilities. After calibrating the rules, Tibo noted they currently favor a reset for users, and carried out the reset the community voted for.
Anthropic Offers Startups a Free Year of Claude Team Service and $1,000 in Credits
Anthropic is rolling out a support program for early-stage startups, which gives participating companies a full year of free access to Claude Team along with $1,000 in model usage credits, according to a TechCrunch report. The company says it built the program because it believes the benefits of AI will reach most people through companies that build products on top of models, rather than through the models themselves.
Tibo’s 28-Day Codex Plan Day 2: Auto-review is now free
Day 2 of Tibo’s 28-day Codex improvement plan brings a concrete change: Auto-review is now free for all users signed in with a ChatGPT account. The feature uses a second agent to review actions taken by the primary agent, reducing approval fatigue from the default sandbox mode while aiming to block high-risk or off-intent actions, with no usage drawn from the user’s plan.
Claude Language Settings Menu Re-Adds Simplified and Traditional Chinese
A shared interface screenshot shows Claude's language selection menu now includes both Simplified Chinese and Traditional Chinese options; American English remains the default selected language, and the screenshot does not state a specific release date or model version.
Tibo’s Day 1 update: GPT-6 Astra and GPT-6.1 Sol get about 50% faster by default
Tibo shared what he called Day 1 of his 28-day improvement push on X: default speed for GPT-6 Astra and GPT-6.1 Sol has been optimized to be about 50% faster across subscription access, including OpenAI products and partners using Sign in With ChatGPT. He said no user action is needed and the change should be felt within two hours.
OpenAI is adding invisible watermarks to eligible ChatGPT and Codex text in the EU
OpenAI has outlined its response to EU AI Act text provenance rules: over the coming weeks, it will add invisible watermarks to eligible ChatGPT and Codex text outputs in the European Union. Global API customers can also opt in to text watermarking for select models, though it stays off by default. The system is called textGrain, but detection remains limited with shorter text, math-heavy content, and edits all reducing reliability.
OpenAI to Test Visual Ads Inside ChatGPT Image Generation
OpenAI plans to start testing a new ad format in the U.S. later this month, showing visual ads while ChatGPT generates images. The ads will be labeled and kept separate from the generated image, with free-tier and Go-tier users affected first.