From Anthropic · · 2m
Claude Fable 5.2 and Mythos 5.2 Launch
Anthropic releases Claude Fable 5.2 and Mythos 5.2 with improved coding, reasoning and vision capabilities.
In brief
Anthropic releases Claude Fable 5.2 and Mythos 5.2 with improved coding, reasoning and vision capabilities. Anthropic launched Claude Fable 5.2 and Mythos 5.2 models with enhanced coding, reasoning and vision features for technical work. Originally reported by Anthropic.
Claude Fable 5.2 and Mythos 5.2 Launch
Anthropic upgraded Claude Fable and Claude Mythos models to tackle complex technical challenges.
Performance Improvements
Claude Fable 5.2 and Mythos 5.2 show improvements in coding, reasoning and vision capabilities.
Prerelease Testing Results
Testers found Fable 5.2 and Mythos 5.2 helped engineers complete software migrations, design electronic devices, solve black hole physics problems, and formalize Fermat’s Last Theorem proof in 30 hours.
Same Underlying Model
Claude Fable and Claude Mythos share the same model with Fable having cybersecurity and biology safeguards for general users.
Access Programs
Organizations can apply for access to Mythos through Anthropic’s Life Sciences and Cyber Verification Programs.
Speed and Cost Improvements
Fable 5.2 and Mythos 5.2 generate outputs up to 30% faster than previous versions at $8 per million input tokens and $40 per million output tokens.
Fast Mode Option
An optional fast mode speeds up token generation by up to 2.5 times at $16 per million input tokens and $80 per million output tokens.
Safety and Alignment
Anthropic ran extensive safety testing showing Fable 5.2’s cyber and biology safeguards are as robust as Fable 5.1 but produce fewer unnecessary blocks.
When to Use
Fable and Mythos are best for ambitious open-ended technical work with Opus 5.5 recommended for most coding and knowledge tasks.
Benchmark Performance
Claude Fable 5.2 and Mythos 5.2 outperform Opus 5.5 in coding, reasoning, cybersecurity and other technical capabilities.
Agentic Coding
Fable 5.2 achieved 49.1% accuracy on FrontierCode 2.0 benchmark at Xhigh effort, the highest score of any released model.
Real-World Applications
Anthropic engineers used Fable 5.2 to migrate a critical service from Python to Rust, making it more efficient, reliable and starting 26 times faster.
Operating System Prototype
Fable 5.2 designed a working operating system from scratch in eight days using over 1,000 subagents.
User Testimonial
Vercel engineer says Fable 5.2 understands prompt intent, completes projects faster and reduces code review comments.
“The model is in my head. It understands the intent behind every prompt I’ve given it. I haven’t had to say ‘No, I meant…’ a single time. Projects we’ve been pushing along with other models for weeks got finished and approved by our teammates in a few days with Claude Fable 5.2, and the humans reviewing its code had less to say.”
At a glance
Anthropic launched Claude Fable 5.2 and Mythos 5.2 models with enhanced coding, reasoning and vision features for technical work.
