Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | Claude 3 Opus | GPT-4.1 mini |
|---|---|---|
| GPQA Diamond | 50.4 | 65 |
| Humanity's Last Exam | — | 3.7 |
math
| Benchmark | Claude 3 Opus | GPT-4.1 mini |
|---|---|---|
| AIME 2025 | — | 40.2 |
coding
| Benchmark | Claude 3 Opus | GPT-4.1 mini |
|---|---|---|
| SWE-bench Verified | — | 23.6 |
multimodal
| Benchmark | Claude 3 Opus | GPT-4.1 mini |
|---|---|---|
| MMMU | — | 72.7 |
| CharXiv Reasoning | — | 56.8 |
general
| Benchmark | Claude 3 Opus | GPT-4.1 mini |
|---|---|---|
| MMMLU | — | 78.5 |
agent
| Benchmark | Claude 3 Opus | GPT-4.1 mini |
|---|---|---|
| TAU-Bench Retail | — | 55.8 |
Category Winners
coding
0.5
math
2.9
reasoning
24.8
general
30.8
multimodal
36.2
agent
18.1