Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | GPT-5 mini | Mistral Small 4 |
|---|---|---|
| GPQA Diamond | 82.3 | 71.2 |
| Humanity's Last Exam | 16.7 | — |
math
| Benchmark | GPT-5 mini | Mistral Small 4 |
|---|---|---|
| AIME 2025 | 91.1 | 83.8 |
| FrontierMath | 22.1 | — |
multimodal
| Benchmark | GPT-5 mini | Mistral Small 4 |
|---|---|---|
| MMMU-Pro | — | 60 |
Category Winners
math
58.8
reasoning
58.9
multimodal
27.7