Claude Fable 5.1 is Anthropic's long-horizon highest-capability public model for long-running agents, hard coding work, and document-heavy research, released on September 1, 2026, and is built for engineering and knowledge teams that push the same large context through many turns. It accepts text and images, returns text, and keeps a 1,000,000-token window with up to 128,000 output tokens.
Core Features
- 1M-token context: it holds an entire repository plus conversation history, so agent loops rarely lose prior state.
- Adaptive thinking: every request reasons by default, with effort settings to trade latency for quality on tough tasks.
- Multimodal input: it reads screenshots and diagrams, not just code, which helps front-end and visual debugging.
- Prompt caching: cache reads dropped 75% to $0.25 per million tokens, so long sessions cost far less than Fable 5.
- Wide availability: reachable through the Claude API, Amazon Bedrock, Google Cloud, and Microsoft Foundry.
Use Cases
- Long code refactors and front-end work where the model must track many files at once.
- Research and document analysis over huge knowledge bases.
- Production agents that re-read the same system prompt every turn.
Pricing
Direct API pricing is $10 per million input tokens and $50 per million output tokens, with cache writes at $12.50 (5-minute) and $20 (1-hour) and cache reads at $0.25. Batch API cuts input and output rates by 50%. A Claude Pro, Max, Team, or Enterprise plan includes Fable 5.1 in the app but not API credits.
Our Take
Best for agentic coding and research where the 1M context earns its keep. The trade-off is premium pricing and latency, so Anthropic suggests starting with Opus 5 for simpler workloads.
Compare it with other models in our AI Engine/Model and AI Programming development categories.
