Updated
AI Pricing This Week: Fable 5.2, Gemini 4 Carbon, Nov 9-15
Short answer. From November 9 to 15, 2026, no list price moved. Six models entered the dataset: Claude Fable 5.2 and Mythos 5.2 at $8 / $40, Muse Ember, Gemini 4 Carbon and Hydrogen, and GPT-6 Rosalind. Gemini 4 Carbon doubles on December 1. OpenAI dated the end of GPT-5.6 Sol’s promotion: November 30.
In plain words. No price changed this week, but six new models arrived. Two come from Anthropic, two from Google, one from Meta and one from OpenAI. Some of them only approved companies can use. One OpenAI discount now has an end date. One Google model will cost twice as much from December.
Update, November 17, 2026: OpenAI cut GPT-6 Rosalind’s price after this week closed. Its pricing page, read in its Markdown version on November 17, 2026, prints “gpt-6-rosalind (short context) | $10.00 | $50.00 | $1.00 | $12.50 | $25.00 |” and “GPT-6 Astra for Law, GPT-6.2 Astra for Law, and GPT-6 Rosalind promotional pricing is available at least through December 31, 2026.” The $15.00 / $75.00 below was the price on the page when we read it on November 9 and still on November 16. The same day OpenAI released GPT-6.2 Astra and GPT-6.2 Sol; the post of November 17 has both.
This is the fifth weekly summary. It covers Monday, November 9 to Sunday, November 15, 2026. Every fact comes from AI Signal’s dataset, change log and deprecation records, with the provider page it was read on and the date. Two items the fourth summary listed as read after its week closed, GPT-5.6 Sol’s promotion and Qwen3.8 Max on Together, are told in full here.
The week in three lines. Anthropic and Google each released two models, Meta one, and OpenAI listed a restricted one. Google’s newest, Gemini 4 Hydrogen, has no end date on its price; Gemini 4 Carbon’s doubles on December 1. Plans and tools kept their prices and changed their wording.
New models
| Model | Released | Added here | Input / output per 1M tokens |
|---|---|---|---|
| Claude Fable 5.2 | November 9 | November 10 | $8.00 / $40.00 |
| Claude Mythos 5.2 | November 9 | November 10 | $8.00 / $40.00, verification required |
| Muse Ember | November 9 | November 10 | $0.10 / $0.50 up to 200K input tokens |
| Gemini 4 Carbon | November 10 | November 11 | $2.00 / $10.00 through November 30, then $4.00 / $20.00 |
| Gemini 4 Hydrogen | November 12 | November 13 | $1.50 / $7.50 |
| GPT-6 Rosalind | no date | November 9 | $15.00 / $75.00, restricted |
Claude Fable 5.2 and Claude Mythos 5.2. Anthropic’s release notes of November 9 say “We’ve launched Claude Fable 5.2 ( claude-fable-5-2 ), built for demanding reasoning and long-horizon agentic work.” and “They cost $8 USD per MTok of input and $40 USD per MTok of output (Claude Fable 5.1 and Claude Mythos 5.1 cost $10 USD per MTok of input and $50 USD per MTok of output).” So Fable 5.1 costs 1.3× more than Fable 5.2 on input and on output. Mythos 5.2 has the same prices but not the same access: “Claude Mythos 5.2 ( claude-mythos-5-2 ) offers the same capabilities only to organizations verified through Anthropic’s verification programs”. Its model page, read on November 10, prints “Status Active (verification required)” and “It has the same specifications and pricing as Claude Fable 5.2.” The full comparison is in the Fable 5.2 post.
Muse Ember. Meta’s model page, read on November 10, prints its two price bands cell by cell: “muse-ember-1.0 <200k context 1M $0.004 $0.10 $0.50” and “muse-ember-1.0 >200k context 1M $0.012 $0.30 $1.50”, that is cached input, input and output per million tokens, below and above 200,000 input tokens. Meta’s pricing page adds “Model: muse-ember-1.0 , in Standard mode only.” It is the only Meta model in the dataset with a price that depends on prompt length.
Gemini 4 Carbon. Google’s changelog reads “November 10, 2026 Gemini 4 Carbon ( gemini-4-carbon ) generally available (GA)”. The pricing page prints the same cells as for Gemini 4 Argon: “$2.00 through November 30, 2026. $4.00 starting December 1, 2026.” for input, and $10.00 then $20.00 for output. It adds “Gemini 4 Carbon is not available on the Free tier.” The Carbon post has the side-by-side with Argon.
Gemini 4 Hydrogen. Two days later the changelog reads “November 12, 2026 Gemini 4 Hydrogen ( gemini-4-hydrogen ) generally available (GA)”, and the pricing page prints “Input price | Free of charge | $1.50 |”, with $7.50 for output and no end date. Carbon costs 1.3× more than Hydrogen on input and output through November 30, and 2.7× more on input from December 1. The Hydrogen post compares it with the other Gemini 4 models.
GPT-6 Rosalind. OpenAI’s pricing page, read on November 9, lists it under “Life sciences models” (“Our latest Rosalind models.”) as “gpt-6-rosalind | $15.00 | $75.00 | $1.50 | $18.75 | $37.50 |”: input, output, cached input and the two cache writes. The block says “Access is limited to approved internal research through the trusted-access program”. Its page here is kept out of the search index, as other restricted models’ pages are.
Gemini Decision 1: no row
Google added Gemini Decision 1 on November 12, and it does not appear in the price table. Its model page prints “Text generation Not supported”, and the pricing page gives its output as “Not applicable (0 output tokens)”, next to $0.10 per million input tokens. The table covers models that are billed per token and write text, so Decision 1 stays out; that is our criterion, recorded with the date, not a judgment of the model. What it costs and what it does are in its own post.
Price changes
No list price moved in the week. One entry of the price-change log is a change of wording by OpenAI, not of price: on November 9 its pricing page and GPT-5.6 Sol’s model page said “GPT-5.6 Sol’s promotional pricing is available until November 30, 2026.”, where every weekly reading from October 5 to November 2 had found “at least through November 21, 2026”. The promotional price, $4.00 / $20.00, is unchanged, and OpenAI still publishes no price for after it.
One increase was announced, not applied: Gemini 4 Carbon’s 2.0× rise on December 1, logged with Argon’s on the same date.
Retired at a host
Qwen3.8 Max on Together AI, removed on November 4. Together’s deprecations page, read on November 9, prints “2026-11-04 | Qwen/Qwen3.8-2.4T-A95B | Qwen/Qwen3.7-Max | No |”: removal date, model, recommended replacement, and “No” for on-demand dedicated endpoints. The weekly check of that day did not find it on Together’s pricing page. The offering stays as a retired record, with its last price, $2.00 / $6.00, on open-model prices by host. The same Monday, Baseten’s pricing page, which a script no longer receives, began to be read with a headless browser; on November 10 its seven prices matched.
Subscriptions and coding tools
Google AI plans name the new models (November 11 and 13). Prices did not move: AI Plus $4.99, AI Pro $19.99 and AI Ultra $99.99 or $199.99 a month. On November 13 the US page prints “Includes Gemini 4 Hydrogen” above AI Plus, “Access to Gemini 4 Hydrogen” where it said “Access to our Flash model” on November 3, and “Includes Gemini 4 Carbon” above AI Ultra. Google Gemini’s plans quote the lines of that day.
Google Antigravity (added November 10). Its pricing page sells one plan of its own, “Generally Available For Individuals $0/month”, and sends Pro and Ultra to the Google AI plans, with no price of its own. Google’s Gemini Code Assist page says “Going forward, agentic coding assistance will be available through Antigravity”, so the two Antigravity and Gemini Code Assist pages link to each other.
Devin: Teams up to 100 users (November 11). On October 19 devin.ai/pricing listed Teams with “Up to 200 users”; on November 11 it lists “Up to 100 users”, and Pro, Max and Teams add “Use your ChatGPT plan”. The Teams price did not move. When the page changed is not known. Devin records the change.
JetBrains AI Free (November 11). Read by hand from Spain, the Free plan adds a line it did not print on October 22: “Only available for commercial IDE versions”. Prices and credits did not move. JetBrains AI quotes it.
GitHub Copilot (November 11). Plan prices did not move. GitHub’s documentation lists a new option, “Copilot access : Available to eligible enterprises.”, which GitHub Copilot quotes but does not count as a plan. Kiro had no dated change in the week; its credit pooling was seen on November 6 and is in the fourth summary.
Posts of the week
- November 9: the fourth weekly summary, November 2 to 8.
- November 10: Claude Fable 5.2.
- November 11: Gemini 4 Carbon.
- November 12: Gemini Decision 1.
- November 13: Gemini 4 Hydrogen.
- November 13: What Changes on December 1, 2026.
What comes next
The December 1 post lists every dated item from November 30 to December 7, all on Deprecations or the price-change log:
- November 30: Anthropic retires Claude Sonnet 4.5: “November 30, 2026 | claude-sonnet-4-5-20250929 | claude-sonnet-5-5 |”. GPT-5.6 Sol’s promotional price ends, with no next price published.
- December 1: Gemini 4 Argon and Gemini 4 Carbon go from $2.00 / $10.00 to $4.00 / $20.00. Gemini 2.5 Pro routes to Gemini 3.8 Flash and Gemini 2.5 Flash to Gemini 3.5 Flash-Lite.
- December 3: Gemini Omni 1.1 Flash routes to Gemini Omni 1.2 Flash.
- December 11: OpenAI shuts down GPT-5, GPT-5 mini, GPT-5 nano, GPT-5 Pro, o3 and o3-pro.
Since the week closed
On Monday, November 16, Alibaba released Qwen4-Max at $1.50 / $4.50 in its Singapore region (the post), and OpenAI’s deprecations page was found to date the shutdown of GPT-4o, GPT-4o mini, GPT-4.1 and GPT-4.1 mini for January 14, 2027, a date AI Signal had missed and has corrected (the post). Both belong to the sixth summary.
How we checked
The facts were listed by our weekly script from AI Signal’s own data files: the models dataset, the price-change log, the deprecation records, the subscription and coding-tool data and the posts of the week. Every price and date was read on the provider’s own page, on the date given; Google’s pages were read from their Markdown source, Baseten’s with a browser, and JetBrains AI’s by hand. Model prices are list prices in US dollars per million tokens; ratios are the larger price divided by the smaller. We do not know why any of these prices are what they are, and this summary does not guess.
Frequently asked questions
Which AI model prices changed between November 9 and 15, 2026?
No list price moved. Six models entered the dataset, among them Claude Fable 5.2 at $8.00 / $40.00 per million tokens and Gemini 4 Hydrogen at $1.50 / $7.50; Google's page says Gemini 4 Carbon goes from $2.00 to $4.00 input on December 1, an announced increase, not one that happened in the week.
What changed in GPT-5.6 Sol's promotion?
Its end date: OpenAI's pricing page, read on November 9, 2026, says 'GPT-5.6 Sol’s promotional pricing is available until November 30, 2026.', where it had said 'at least through November 21, 2026'. The promotional price, $4.00 / $20.00, did not change, and OpenAI publishes no price for after it.
Did any AI subscription or coding tool change between November 9 and 15, 2026?
No subscription or tool price moved. Google's US plans page named Gemini 4 Hydrogen on November 13, Devin's Teams plan listed 'Up to 100 users' on November 11 where it had listed 200, JetBrains AI Free added 'Only available for commercial IDE versions', and Google Antigravity entered the tools pages.
Written with AI assistance from this site's own verified pricing dataset and reviewed by Álvaro Lucero Rufino, who is responsible for what stays published here. Every figure links to the page where it was verified, with the date it was read; if a number here and a number on that page ever disagree, the page is the one to trust. Automated publishing was discontinued in September 2026; posts that no longer met that standard were removed rather than left up.