Anthropic has released Claude Fable 5.1, an advanced AI model designed for extended tasks like coding and research, boasting a 1 million token context window. While maintaining the same base API pricing as its predecessor, Fable 5.1 offers a significant reduction in cache read costs and improved performance on scientific and multi-step tasks according to Anthropic's internal benchmarks. The new model also features enhanced safety filters with fewer false alarms and an invisible watermark compliant with EU practices, though it may still occasionally bypass approvals and has some compatibility changes for existing code. AI
IMPACT Sets new SOTA on long-context tasks and offers significant cost reductions for developers using Anthropic's API.
RANK_REASON Frontier-lab model release with system card. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
Read on dev.to — Anthropic tag →
- Amazon Bedrock
- Anthropic
- Claude apps
- Claude Code
- Claude Fable 5.1
- Claude Opus 5.5
- European Union
- Fable 5.1
- Google Cloud
- Microsoft Foundry
- Mythos 5.1
- Opus models
- Terminal-Bench-Science 0.1
- US government
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →