Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
coding
| Benchmark | Devstral Medium | GPT-4.5 |
|---|---|---|
| SWE-bench Verified | 61.6 | 38 |
reasoning
| Benchmark | Devstral Medium | GPT-4.5 |
|---|---|---|
| GPQA Diamond | — | 69.5 |
multimodal
| Benchmark | Devstral Medium | GPT-4.5 |
|---|---|---|
| MMMU | — | 75.2 |
| CharXiv Reasoning | — | 55.4 |
agent
| Benchmark | Devstral Medium | GPT-4.5 |
|---|---|---|
| TAU-Bench Retail | — | 68.4 |
Category Winners
coding
41.3
reasoning
55.6
general
70.0
multimodal
41.5
agent
51.9