Claude Fable 5
Anthropic's most capable widely released model, for demanding reasoning and long-horizon agentic work.
Intelligence
AA Intelligence Index60
independent benchmark score from Artificial Analysis, not Timace's own judgment
Best for
- ✓ Long-horizon agentic work and the hardest reasoning tasks
Not ideal for
- – High-volume, latency-sensitive workloads
Specifications
- Developer
- Anthropic
- Context window (tokens)
- 1M tokens
- Input $/MTok
- $10 / MTok
- Output $/MTok
- $50 / MTok
- Modalities
- text, image, code
- API available
- Yes
- Open source
- No
