Claude Fable 5 vs GPT-5.6 Sol
Claude Fable 5 (Anthropic) and GPT-5.6 Sol (OpenAI) compared on price, context window, speed and independent benchmarks. Claude Fable 5 scores higher on the Artificial Analysis Intelligence Index (49.6 vs 47), while GPT-5.6 Sol is 2.5× cheaper on a blended basis. Pick Claude Fable 5 for the hardest tasks and GPT-5.6 Sol when cost dominates.
| Claude Fable 5 | GPT-5.6 Sol | |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Input $/1M | $10.0 | $4.00 |
| Output $/1M | $50.0 | $20.0 |
| Blended $/1M (3:1) | $20.0 | $8.00 |
| Cached input $/1M | $1.00 | $0.40 |
| Context window | 1M | 1.1M |
| Knowledge cutoff | 2026-01 | 2026-02 |
| License | proprietary | proprietary |
| Speed (tok/s) | 68 | 58 |
| Artificial Analysis Intelligence Index | 49.6 | 47 |
| LMArena Elo | 1493 | 1455 |
| GPQA Diamond | 92.6% | 94.1% |
| AIME | 99.7%* | — |
| SWE-bench Verified | 95% | — |
| MMLU-Pro | 91.5%* | — |
| Humanity's Last Exam | 55.5% | 49.5% |