Anthropic releases Claude Fable 5.1 and Mythos 5.1 as new flagship models with updated pricing and benchmark improvements

Anthropic announced the release of Claude Fable 5.1 and Claude Mythos 5.1, describing them as the world’s most advanced models for coding and knowledge work. The company framed Fable 5.1 specifically for autonomous, multi-step tasks such as coding, knowledge work, and long-running problem solving.

Pricing for Fable 5.1 remained unchanged from Fable 5 for input tokens ($10 per million tokens), output tokens ($50 per million tokens), and cache write ($12.5 per million tokens), but cache read pricing was reduced by 75% to $0.25 per million tokens. Anthropic highlighted this change as beneficial for long sessions and long-context use cases, particularly for agentic workloads where prompts are frequently re-read from cache.

Early benchmark results reported by Artificial Analysis indicated improvements across several evaluations. On the Artificial Analysis Intelligence Index, Fable 5.1 achieved a score of 66 at max effort, surpassing Claude Opus 5 (63), Claude Fable 5 (62), and GPT-5.6 Sol (61). Additional gains were reported on Terminal-Bench-Science (52.6%), SciCode (62.0%), and τ³-Banking (+9 points over Fable 5).

However, community analysis noted that Fable 5.1’s per-task cost increased by approximately 20% compared to Fable 5, driven by a 1.7x increase in output token usage. Artificial Analysis estimated Fable 5.1’s cost per task at $3.76, higher than Fable 5’s lower cost, though the cache price cut partially offset this increase.

Anthropic also introduced new enterprise-oriented controls, including Enterprise Frontier Safeguards (EFS), described as an enhancement for agent observability in enterprise environments. The company emphasized zero-data-retention support as an adoption enabler.

User and community reactions highlighted both praise for coding and planning capabilities and concerns about rate limits, safeguard false positives, subscription user experience, and benchmark presentation clarity.