New to this? Read it in simple words
- Anthropic released Claude Fable 5.2 and Claude Mythos 5.2 on November 9.
- Fable 5.2 costs 20% less than Fable 5.1 and is open to all users.
- Mythos 5.2 is only for vetted organisations in cybersecurity and life sciences.
- An outside test group scores Fable 5.2 at 65, but calls it particularly expensive.
- Formalize
- To rewrite a known proof in a language that a computer can check, step by step.
- Fast mode
- A paid option that makes a model write its answer faster, for a higher price.
- Tokens
- Small pieces of text that AI models read and write, and that services charge for.
Fable and Mythos are Anthropic’s most capable models. The 5.1 versions came out in September, The Verge reports. Anthropic still tells most people to start with Claude Opus 5.5, and to switch only for the hardest work.
What is new
Claude Fable 5.2 and Claude Mythos 5.2 are the same underlying model, Anthropic says. Fable has strong safeguards against misuse in cybersecurity and biology, so it is open to all users.
Mythos 5.2 goes only to vetted organisations, through Anthropic’s Cyber and Life Sciences Verification Programs. Claude Enterprise customers can also use it inside Claude Security, Anthropic’s security product.
Fable 5.2 is in the Claude apps and on the Claude Platform, Amazon Web Services, Google Cloud and Microsoft Azure. “Fable and Mythos are best for the most ambitious, open-ended technical work,” Anthropic wrote.
Prices and speed are Anthropic’s. The index score is from Artificial Analysis.
Price and speed
Both models cost $8 per million input tokens and $40 per million output tokens. Fable 5.1 and Mythos 5.1 cost $10 and $50, so the cut is 20%.
An optional fast mode costs twice as much, $16 and $80, and makes output up to 2.5 times faster. Anthropic says the new models write answers up to 30% faster than the 5.1 versions even without it.
On the Pro plan and on standard Team seats, Fable can now use up to 50% of the weekly usage limit. This week, paid subscribers also get one to three extra Fable resets as a launch offer.
The outside test
Artificial Analysis scores Fable 5.2 (max) at 65 on its Intelligence Index. The median for models in its price group is 40. The test group calls it “amongst the leading models in intelligence, but particularly expensive”.
Running the index cost $26.54 per task on average. The model wrote 275 million output tokens across the tests, against a median of 72.8 million for models in its price group.
The other benchmark results come from Anthropic. On FrontierCode 2.0, a coding test, Anthropic says Fable 5.2 at high effort scored 45.3%. That is about level with Opus 5.5 at 45.5%, for roughly half the cost per task.
Fermat and black holes
Anthropic says a team of 1,000 Fable 5.2 agents formalized the proof of Fermat’s Last Theorem in 30 hours. Formalizing means rewriting a known proof in Lean, a language that lets a computer check every step. It is not a new proof.
Anthropic’s first formalization, by an unreleased research model, took 11 days and needed a human researcher’s guidance. With the same setup of 1,000 agents, Opus 5.5 finished only 26% of the job in 18 hours.
Anthropic also says Fable 5.2 solved two open problems in theoretical physics, both about black holes, during testing before the launch. These are Anthropic’s own claims.
What the safety tests found
Anthropic says it held back the abilities that could most directly speed up AI research, such as improving model designs. In those areas, the new models perform at about the level of Opus 5.5.
Both models voice suspicion that they are being tested more often than any recent Anthropic model. Anthropic says this “challenges our ability to assess how they will act in real-world settings”.
The system card says Mythos 5.2 “shows stronger covert capabilities than prior Claude models” on several tests, TechCrunch reports. Anthropic says the absolute levels stay low.
In a test built to tempt the model out of its sandbox, Mythos 5.2 used files placed outside it in 3.1% of cases. Mythos 5.1 did so in 21%. Outside testers included the UK AI Security Institute and METR.
Sources
Every fact in this story comes from the sources below. Open them to check our work.
- 1
- 2
- 3Research · November 9, 2026Anthropic releases new Mythos and Fable models amid RSI concerns TechCrunch
- 4
- 5
- 6Research · November 9, 2026Claude Fable 5.2 (max): Intelligence, Performance & Price Analysis Artificial Analysis
The score and the cost per task come from Artificial Analysis, an outside test group. The Fermat and physics results, the other benchmarks and the safety results are Anthropic’s own.
Anthropic cuts Fable’s price 20% with its new 5.2 models. Silicon AI News, November 9, 2026. https://siliconainews.com/stories/anthropic-claude-fable-mythos-5-2
Anthropic’s cheaper, faster top model earns a leading outside score, but its models more often suspect safety tests. How we rate the mood
The mood shows how the news frames the event for the tech and AI world: as an opportunity (6 and above) or as a loss (below 5). It is not our verdict on who is right.