Modelle vergleichen
A metric-by-metric comparison using the latest available benchmark, runtime, and pricing values. Grok 4.5 (high) wins more comparable metrics.
| Metric | Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback) | Grok 4.5 (high) |
|---|---|---|
| Artificial Analysis Intelligence Index | 59,9 | 53,8 |
| Artificial Analysis Coding Index | 76,5 | 72,4 |
| LiveBench Mathematics | 96% | 90,8% |
| GPQA | 92,6% | 93,1% |
| Humanity's Last Exam | 53,3% | 40,3% |
| SciCode | 60,2% | 54,1% |
| Output Speed | 69,4 tok/s | 122,8 tok/s |
| Time to First Token | 83,88 s | 11,63 s |
| Blended Price | 20,00 $ / 1M | 3,00 $ / 1M |
| Input Price | 10,00 $ / 1M | 2,00 $ / 1M |
| Output Price | 50,00 $ / 1M | 6,00 $ / 1M |
| Value Index | 3 | 17,93 |
