Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | GPT-5 mini | Mercury 2 |
|---|---|---|
| GPQA Diamond | 82.3 | 74 |
| Humanity's Last Exam | 16.7 | — |
math
| Benchmark | GPT-5 mini | Mercury 2 |
|---|---|---|
| AIME 2025 | 91.1 | 91.1 |
| FrontierMath | 22.1 | — |
coding
| Benchmark | GPT-5 mini | Mercury 2 |
|---|---|---|
| SciCode | — | 38 |
Category Winners
coding
I
Mercury 218.9
math
I
Mercury 271.6
reasoning
I
Mercury 264.3