← All comparisons
Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) vs Grok 4.20 0309 v2 (Non-reasoning)
Anthropic vs xAI — side-by-side benchmark comparison
| Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) | Grok 4.20 0309 v2 (Non-reasoning) | |
|---|---|---|
| Intelligence Index | 51.7 | 29.0 |
| Coding Index | 50.9 | 22.0 |
| Math Index | — | — |
| Output speed (tok/s) | 68.2 | 175.2 |
| Blended price ($/1M) | $6.56 | $3.00 |
| Time to first token (s) | 55.35s | 0.47s |
| aime | — | — |
| aime 25 | — | — |
| artificial analysis coding index | 50.90 | 22.00 |
| artificial analysis intelligence index | 51.70 | 29.00 |
| artificial analysis math index | — | — |
| gpqa | 87.5% | 77.6% |
| hle | 30.0% | 24.2% |
| ifbench | 56.6% | 49.3% |
| lcr | 70.7% | 17.3% |
| livecodebench | — | — |
| math 500 | — | — |
| mmlu pro | — | — |
| scicode | 46.8% | 32.8% |
| tau2 | 75.7% | 59.9% |
| terminalbench hard | 53.0% | 16.7% |
Benchmark data from Artificial Analysis.