← All comparisons
Llama 3.3 Instruct 70B vs Grok 4.3 (low)
Meta vs xAI — side-by-side benchmark comparison
| Llama 3.3 Instruct 70B | Grok 4.3 (low) | |
|---|---|---|
| Intelligence Index | 14.5 | 43.9 |
| Coding Index | 10.7 | 31.6 |
| Math Index | 7.7 | — |
| Output speed (tok/s) | 88.1 | 118.3 |
| Blended price ($/1M) | $0.62 | $1.56 |
| Time to first token (s) | 0.59s | 6.35s |
| aime | 30.0% | — |
| aime 25 | 7.7% | — |
| artificial analysis coding index | 10.70 | 31.60 |
| artificial analysis intelligence index | 14.50 | 43.90 |
| artificial analysis math index | 7.70 | — |
| gpqa | 49.8% | 84.3% |
| hle | 4.0% | 17.3% |
| ifbench | 47.1% | 81.0% |
| lcr | 15.0% | 64.0% |
| livecodebench | 28.8% | — |
| math 500 | 77.3% | — |
| mmlu pro | 71.3% | — |
| scicode | 26.0% | 41.9% |
| tau2 | 26.6% | 88.9% |
| terminalbench hard | 3.0% | 26.5% |
Benchmark data from Artificial Analysis.