OpenRouter lists Claude Fable 5.1 with a Sep 1, 2026 release date, 1M context, and listed pricing of $10/$50 per 1M tokens with $0.25 cache reads, routing requests across Anthropic, Google Vertex, Amazon Bedrock (BYOK only), and Azure. The Azure endpoint shows $10/$50 listed pricing with $0.25 cache reads, 6.49s latenc Effective weighted-average pricing across providers sits at $1.802/M input tokens and $49.99/M output, driven heavily by Anthropic's 84.3% cache hit rate producing an effective $1.989/M input versus Vertex's $1.095/M at a 73.1% cache hit rate and Bedrock's $1.005/M at 93.8%. Bedrock's endpoint is BYOK-only pending publ

Anthropic announced Claude Fable 5.1 and Claude Mythos 5.1 on September 1, 2026 as configurations of the same underlying model with different safeguards: Fable 5.1 is generally available, while Mythos 5.1 is restricted to Anthropic's trusted-access programs for cybersecurity and life-sciences work. Fable 5.1 is availab Anthropic's published benchmarks for Fable 5.1 include 52.6% on Terminal-Bench-Science 0.1 (vs 24.7% for Fable 5, 29.0% for Opus 5), 55.8% on Terminal-Bench 4.0 (Mythos 5.1 reaches 60.9%), 73.4% on CursorBench 3.2.0, 60.9% without tools and 65.0% with tools on Humanity's Last Exam, and a GDPval-AA v2 knowledge-work sco

Data Science Dojo's September 1, 2026 analysis frames Fable 5.1 as a targeted push on two fronts enterprise and research teams have been asking for: hard-task performance and safety systems that stop blocking legitimate work. Concrete benchmark jumps include more than doubling Terminal-Bench-Science from 24.7% to 52.6% On developer-facing behavior, Fable 5.1 ships always-on adaptive thinking that replaces the opt-in extended thinking from prior models, with an effort parameter that lets teams trade latency for reasoning depth. Pricing is reported as roughly 25% lower than Fable 5 overall and up to 45% lower on heavily agentic workloa

The Releasebot-curated feed confirms the September 1, 2026 launch of Claude Fable 5.1 and Claude Mythos 5.1 as Anthropic's most advanced models for coding and knowledge work, pointing to Anthropic's first-party blog post and Developer Platform release note. The Developer Platform entry mirrors the official release note The same feed captures an unrelated Claude Code fix shipped the same day: Claude Code 2.1.258 resolves a macOS 12 (Monterey) launch regression introduced in 2.1.255, and fixes remote and scheduled sessions failing with "user messages must have non-empty content" after a re-sent permission approval could not be applied.

On September 1, 2026, Anthropic released Claude Fable 5.1 (model ID claude-fable-5-1) as the successor to Fable 5 for long-running agentic coding, knowledge work, and research, alongside Claude Mythos 5.1 for Project Glasswing participants. Both ship with a 1M token default context window, 128k max output tokens, and a Several API-breaking changes require developer attention on Fable 5.1 and Mythos 5.1: tool-choice types any and tool return a 400 error (auto and none remain valid, with strict tool use or structured outputs needed to guarantee schema-conformant inputs); thinking-block preservation is now model-scoped (Fable 5.1 accept