Comparer les modèles
A metric-by-metric comparison using the latest available benchmark, runtime, and pricing values. Tie wins more comparable metrics.
| Metric | Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback) | GLM-5.2 (max) |
|---|---|---|
| Artificial Analysis Intelligence Index | 59,9 | 51,1 |
| Artificial Analysis Coding Index | 76,5 | 68,8 |
| LiveBench Mathematics | 96% | 89,8% |
| GPQA | 92,6% | 89,5% |
| Humanity's Last Exam | 53,3% | 40,1% |
| SciCode | 60,2% | 50,5% |
| Output Speed | 69,4 tok/s | 149,2 tok/s |
| Time to First Token | 83,88 s | 0,86 s |
| Blended Price | 20,00 $US / 1M | 2,15 $US / 1M |
| Input Price | 10,00 $US / 1M | 1,40 $US / 1M |
| Output Price | 50,00 $US / 1M | 4,40 $US / 1M |
| Value Index | 3 | 23,77 |
