← All comparisons
GPT-5.4 mini (Non-Reasoning) vs Ministral 3 14B
OpenAI vs Mistral — side-by-side benchmark comparison
| GPT-5.4 mini (Non-Reasoning) | Ministral 3 14B | |
|---|---|---|
| Intelligence Index | 23.3 | 16.0 |
| Coding Index | 25.3 | 10.9 |
| Math Index | — | 30.0 |
| Output speed (tok/s) | 166.4 | 90.6 |
| Blended price ($/1M) | $1.69 | $0.20 |
| Time to first token (s) | 0.48s | 0.37s |
| aime | — | — |
| aime 25 | — | 30.0% |
| artificial analysis coding index | 25.30 | 10.90 |
| artificial analysis intelligence index | 23.30 | 16.00 |
| artificial analysis math index | — | 30.00 |
| gpqa | 60.6% | 57.2% |
| hle | 5.7% | 4.6% |
| ifbench | 38.8% | 32.0% |
| lcr | 30.3% | 22.0% |
| livecodebench | — | 35.1% |
| math 500 | — | — |
| mmlu pro | — | 69.3% |
| scicode | 39.6% | 23.6% |
| tau2 | 23.4% | 27.2% |
| terminalbench hard | 18.2% | 4.5% |
Benchmark data from Artificial Analysis.