Claude Fable 5
used in
0
stacks
Anthropic's most capable widely released model, built for demanding reasoning and long-horizon agentic work.
Fable 5 (API ID claude-fable-5) is described by Anthropic as a Mythos-class model made safe for general use, generally available since June 9, 2026 on the Claude API, AWS, Amazon Bedrock, Vertex AI, and Microsoft Foundry. It has a 1M-token context window by default, up to 128k output tokens, and adaptive thinking that is always on (raw chain of thought is never returned). Anthropic states it can work autonomously for longer than any previous Claude model and supports the memory tool, code execution, programmatic tool calling, context editing, and compaction. It includes safety classifiers that can decline requests, with API-level fallback and fallback-credit billing for retries on other Claude models.
Strengths
- Vendor-stated state-of-the-art on nearly all tested benchmarks, with the lead growing on longer, more complex tasks
- 1M-token context window and up to 128k output tokens per request
- Built for long-horizon autonomous agent work with memory tool, compaction, and programmatic tool calling
- Server-side fallback parameter and fallback credit simplify refusal handling
Tradeoffs
- Premium pricing: $10 per million input tokens and $50 per million output tokens
- Safety classifiers can refuse requests (vendor says under 5% of sessions on average, sometimes catching harmless ones), so integrations must handle stop_reason refusal
- 30-day data retention required; not available under zero data retention