New AI models Claude Fable 5.2 and Mythos 5.2 are now easier to use on the command line
Claude Fable 5.2 is faster and cheaper than before. A small tool update lets anyone type prompts directly instead of clicking around a website.

How it works
A pair of AI models for coding and science tasks, accessible through a command-line tool.
- Install the free llm-anthropic tool
- Get a free API key from Anthropic
- Type a prompt with the model name
- Receive the model's response instantly
What you could do with it
- Engineers could complete complex software migrations that were previously considered impossible, saving time on major tech upgrades 3.
- Researchers could solve difficult physics problems, such as the two open problems in black hole physics that testers used this model for 3.
- Developers could generate working designs for electronic devices or formalize complex mathematical proofs directly from command-line prompts 3.
What changed
Anthropic, the company behind Claude, released Claude Fable 5.2 and Claude Mythos 5.2 on November 9, 2026 3. They are built for tough technical jobs like writing code, solving physics problems, and designing electronics from scratch .
Simon Willison, a well-known developer, updated his open-source tool llm-anthropic to version 0.31 so users can access both models through simple commands 12.
With the tool, anyone can install it and start testing by typing something like:
llm -m claude-fable-5.2 "Generate an SVG of a pelican riding a bicycle"
Willison tested the model at five different reasoning levels — low, medium, high, xhigh, and max — and found that the highest setting, "max," sometimes crashes by running out of tokens after thinking for too long 1. This same problem has appeared with other advanced models like Opus 5.5 .
What you can do with Fable 5.2
According to Anthropic, Fable 5.2 and its sibling Mythos 5.2 helped engineers complete software migrations that previously seemed impossible, designed three working electronic devices, solved two open problems in black hole physics, made progress on robotics tests, and even formalized the proof of Fermat's Last Theorem in 30 hours 3.
Both models produce outputs up to 30% faster than the previous Fable 5.1 and Mythos 5.1 versions 3. They also come with an optional fast mode that can speed things up by 2.5 times when needed .
How much does it cost?
The regular pricing for Fable 5.2 and Mythos 5.2 is $8 per million input tokens and $40 per million output tokens 3. Fast mode costs double: $16 for input and $80 for output .
Mythos 5.2 is aimed at organizations doing beneficial work in cybersecurity or biomedical research. Access requires applying through Anthropic's Life Sciences and Cyber Verification Programs 3. Fable 5.2 has stronger safety guards around cybersecurity and biology topics, which is why it is available to everyone .
Why the tool matters
Open-source tools like llm-anthropic give people a simpler way to experiment with AI without paying for subscriptions or navigating websites. You just need to get a free API key from Anthropic and run a few commands 1. For developers, students, and hobbyists who want to test how modern AI handles coding, science questions, or creative tasks, this means less friction and lower costs than using commercial platforms.
How good is it
Anthropic reports that Fable 5.2 and Mythos 5.2 produce outputs up to 30% faster than the previous Fable 5.1 and Mythos 5.1 versions 3. An optional fast mode can speed things up by 2.5 times when needed . Anthropic says the models performed as well as or better than their recent models on most alignment and safety measures during testing . There are no published benchmark scores against rival models in the sources.
Corrections and updates
- Corrected on 10 November 2026. An earlier version claimed that people previously needed a website or separate program to access the models; none of the source material supports that prior-access assertion, nor does it confirm it should simply be removed.
Key facts
- Claude Fable 5.2 and Mythos 5.2 were released by Anthropic on November 9, 2026
- Output is up to 30% faster and costs $8 per million input / $40 per million output tokens
- llm-anthropic v0.31 lets anyone run these models from the command line
- Fast mode available for double the price with up to 2.5x speed boost
Sources
Keep reading
New modelsAnthropic upgrades Claude Fable and Mythos to version 5.2 with faster speeds and lower pricesHerald of AINew modelsAnthropic upgrades Claude Fable and Mythos to version 5.2 with faster speeds and lower prices
Claude Fable 5.2 and Mythos 5.2 are now available — faster, cheaper, and stronger at coding and hard reasoning tasks than before.






