Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | GPT-5 mini | Qwen3 Max |
|---|---|---|
| GPQA Diamond | 82.3 | 62 |
| Humanity's Last Exam | 16.7 | — |
math
| Benchmark | GPT-5 mini | Qwen3 Max |
|---|---|---|
| AIME 2025 | 91.1 | 81.6 |
| FrontierMath | 22.1 | — |
coding
| Benchmark | GPT-5 mini | Qwen3 Max |
|---|---|---|
| SWE-bench Verified | — | 69.6 |
Category Winners
coding
Q
Qwen3 Max59.1
math
Q
Qwen3 Max54.5
reasoning
48.0