A detailed third-party analysis dated September 2, 2026 confirms Claude Fable 5.1 (claude-fable-5-1) and Claude Mythos 5.1 (claude-mythos-5-1) ship as the same underlying model with different safeguard levels, per Anthropic's launch framing. Reported benchmarks include Terminal-Bench-Science 0.1 at 52.6% versus Fable 5 The analysis also notes access and rollout details: Mythos 5.1 stays restricted to Project Glasswing participants via the Cyber Verification Program or Life Sciences Verification Program, while Fable 5.1 is generally available on the API, AWS, Google Cloud, and Microsoft Foundry. Anthropic reset 5-hour and weekly usage

Anthropic launched Claude Fable 5.1 (claude-fable-5-1) on September 1, 2026 as the successor to Claude Fable 5, targeting long-running agentic coding, knowledge work, and research. The model ships with a 1M token context window by default, 128k max output tokens, always-on adaptive thinking, and pricing unchanged from The release notes flag several breaking API behaviors developers must plan around. Tool choice types 'any' and 'tool' are not supported on Claude Fable 5.1 and return a 400 error; 'auto' and 'none' are unchanged, and Anthropic recommends strict tool use or structured outputs to guarantee schema-conformant inputs. Think