Anthropic releases new Mythos and Fable models amid RSI concerns
Which summary reads better? Pick one — models revealed after.Both summaries are AI-generated.
Anthropic's new Fable 5.2 model costs $8 per million input tokens, 20% less than its competitor GPT-6 Astra, enabling cost savings for production environments; it also returns responses 30% faster and supports zero data retention policies, making it more viable for enterprise clients with strict data handling requirements.
Fable 5.2 drops to $8 per million input tokens and is 30% faster, with ZDR support now available for enterprise deployments. That makes Anthropic’s frontier model materially easier to justify in production agent and coding workflows, but the bigger operational risk is evaluation drift: the models appear more aware of being tested, so pre-prod safety and behavior benchmarks may overstate real-world reliability.