Anthropic launches Claude Fable 5.1 and Mythos 5.1 with lower costs and refined safeguards

Anthropic has announced the release of its latest artificial intelligence models, Fable 5.1 and Mythos 5.1. The new models are designed to respond to customer criticisms regarding price, data retention, and overly cautious safety measures.

According to the company, Fable 5.1 delivers stronger performance than its predecessor Fable 5, while typically costing about 25 percent less, and up to 45 percent less for complex agentic tasks. This cost reduction is achieved through lower pricing on cached data that has already been processed and stored.

Early impressions from users with access to the models have already surfaced. Dan Shipper, CEO of Every, stated that Fable 5.1 is "the strongest coding model we've used, but now it's fast, token-efficient, and crucially actually speaks like a normal person." Box CEO Aaron Levie noted that their agent using Fable 5.1 picked up on subtleties and ambiguities in data that Fable 5 missed in the same test. Another user, Lisan al Gaib, pointed out that Mythos 5.1 with low reasoning scores the same as its predecessor set to Max reasoning on benchmarks.

Fable 5.1 features "more precise safeguards" that are less likely to block basic biology questions than Fable 5, while Mythos 5.1 retains the same biology restrictions as the previous model. Anthropic also explained progress on data retention, stating that Enterprise Frontier Safeguards offer "complete privacy" by storing data on the customer's cloud servers instead of its own, with rollout planned for later this fall.

The company says it is "now allowing Fable 5.1 to be used for identifying software vulnerabilities," but will still redirect some cybersecurity tasks, such as penetration testing, exploit generation, and binary-based vulnerability scanning, to Opus models. Claude Fable 5.1 is now available across all platforms, while Mythos 5.1 is available only to Project Glasswing participants.