Best for agents
Claude Fable 5 is Anthropic's Mythos-class flagship, built for long-horizon autonomous work: multi-file refactors, research synthesis and agentic pipelines that run for hours. Reasoning is always on and the effort level is tunable from low to max, so you can trade latency for depth per request. It accepts text, images and documents across a one-million-token context window.
Extended reasoning
Tool calling
Structured outputs
Vision
Document input
Streaming
Prompt caching
| Model ID | claude-fable-5 |
| Context window | 1,000,000 tokens |
| Max output | 128,000 tokens |
| Inputs | text, image, file |
- Input
- 1,000 cr / M
- Output
- 5,000 cr / M
Best value
GPT-5.6 Sol is the flagship of OpenAI's GPT-5.6 series and the strongest general-purpose choice for command-line work, multi-step coding and tool-driven agents. Reasoning effort scales from none to max, so the same model serves both low-latency chat and deep analysis. It handles text, images and files across a 1.05-million-token context window.
Adjustable reasoning
Tool calling
Structured outputs
Vision
Document input
Streaming
Prompt caching
Seed support
| Model ID | gpt-5.6-sol |
| Context window | 1,050,000 tokens |
| Max output | 128,000 tokens |
| Inputs | text, image, file |
| Knowledge cutoff | 2026-02-16 |
- Input
- 500 cr / M
- Output
- 3,000 cr / M
