Skip to content

The Durability Curve · free by email

Every new piece as it ships, and the Weekly on Saturday, with the calls we grade.

108Pieces, dated
1245Claims verified at the primary source
7Instruments, no login

Your address goes to Substack, where you confirm it. Nothing is stored on this site.

The latest issue · Claude May Have Helped Prove the Riemann Hypothesis. NVIDIA Has Promised to Cover Up to $118 Billion. Read it first

AI model releases in November 2026

Nine releases so far from five labs, six of them frontier models plotted on the release chart. Every lab’s own release log has been read through 20 November 2026.

AI model releases in November 2026, newest first: the day (or the month, where the day was not checked), the lab, the model, what changed and the lab's own post.
When Lab Model What changed
Google Gemini 4 Neon Sits between Hydrogen and Carbon, for software engineering, knowledge work and science. Generally available on launch day on the paid Gemini API, not the free tier, and rolling out the same day in the Gemini app for Google AI Pro and Ultra. $3.00 in and $15.00 out blog.google
OpenAI GPT-6.2 Astra / Sol Astra for the hardest work and Sol for everyday work, released together in the API and, the same day, rolling out in ChatGPT Work and Codex for Plus and higher plans (not regular Chat). Astra lists 20% below GPT-6 Astra on input and output, and Sol keeps GPT-6.1 Sol’s price developers.openai.com
Alibaba Qwen4-Max Flagship multimodal model on ordinary pay-as-you-go API pricing in Model Studio, international and China mainland, five days after the Token Plan-only preview alibabacloud.com
Google Gemini 4 Hydrogen Fast tier and the successor to 3.8 Flash, generally available on launch day on every tier of the Gemini API and AI Studio, the free tier included, in Search for all users and in the Gemini app on paid plans; free app users get it only when auto model selection picks it. $1.50 in and $7.50 out on the paid API blog.google
Alibaba Qwen4-Max-Preview not on the chart Flagship multimodal preview; on the API only through the Token Plan subscription from Alibaba Cloud, with no pay-as-you-go price, and in early access in QwenWork, Qoder, the Qwen app and Qwen Studio; the preview runs from 11 to 25 November 2026 (UTC+8), after which the model is retired alibabacloud.com
Google Gemini 4 Carbon The most capable Gemini model, generally available on the paid Gemini API and Gemini for business and rolling out to Google AI Ultra subscribers in the Gemini app and AI Mode in Search; Fairwind Program cyber defenders had it from the week of 28 Oct blog.google
Anthropic Claude Mythos 5.2 not on the chart Restricted sibling of Fable 5.2, the same model without certain of its cyber and biology safeguards, for organisations in the Cyber and Life Sciences Verification Programs anthropic.com
Anthropic Claude Fable 5.2 Generally available on every platform from launch, at $8 in and $40 out, below Fable 5.1; its sibling Mythos 5.2 is for vetted organisations only anthropic.com
Meta Muse Ember not on the chart Smallest and most efficient Muse model, for agent subtasks such as tool calls, coding and extraction; in Muse Code and on the Meta Model API at $0.10 in and $0.50 out for prompts up to 200K tokens research.meta.ai

A model marked not on the chart is in the full log but not plotted: a smaller tier, a release open only to partners or inside one product, or an image, OCR or research model. The release chart plots the frontier models anyone could use on the day given, public previews and betas included.

See every release on the chart

Every month in the record

The newsletter · Free

Subscribe to get each new piece in your inbox, with something to use in every one.

Get the Weekly: the week in AI, with dated calls graded in public.

Several essays a week, plus Notes through the week. See recent issues

Signup completes on Substack. Unsubscribe from any email.