← All comparisons
Apertus 70B Instruct vs GPT-5.2 (xhigh)
Swiss AI Initiative vs OpenAI — side-by-side benchmark comparison
| Apertus 70B Instruct | GPT-5.2 (xhigh) | |
|---|---|---|
| Intelligence Index | 7.7 | 51.3 |
| Coding Index | 1.9 | 48.7 |
| Math Index | — | 99.0 |
| Output speed (tok/s) | 0.0 | 77.2 |
| Blended price ($/1M) | $1.34 | $4.81 |
| Time to first token (s) | 0.00s | 64.17s |
| aime | — | — |
| aime 25 | — | 99.0% |
| artificial analysis coding index | 1.90 | 48.70 |
| artificial analysis intelligence index | 7.70 | 51.30 |
| artificial analysis math index | — | 99.00 |
| gpqa | 27.2% | 90.3% |
| hle | 5.5% | 35.4% |
| ifbench | 25.9% | 75.4% |
| lcr | 0.0% | 72.7% |
| livecodebench | — | 88.9% |
| math 500 | — | — |
| mmlu pro | — | 87.4% |
| scicode | 5.7% | 52.1% |
| tau2 | 12.9% | 84.8% |
| terminalbench hard | 0.0% | 47.0% |
Benchmark data from Artificial Analysis.