Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | GPT-5.6 Luna | Grok-3 Mini |
|---|---|---|
| GPQA Diamond | 92.3 | 84 |
coding
| Benchmark | GPT-5.6 Luna | Grok-3 Mini |
|---|---|---|
| SWE-Bench Pro | 62.7 | — |
multimodal
| Benchmark | GPT-5.6 Luna | Grok-3 Mini |
|---|---|---|
| MMMU-Pro | 78.4 | — |
agent
| Benchmark | GPT-5.6 Luna | Grok-3 Mini |
|---|---|---|
| BrowseComp | 83.3 | — |
| Toolathlon | 53.4 | — |
math
| Benchmark | GPT-5.6 Luna | Grok-3 Mini |
|---|---|---|
| FrontierMath | 78.6 | — |
| AIME 2025 | — | 90.8 |
general
| Benchmark | GPT-5.6 Luna | Grok-3 Mini |
|---|---|---|
| MRCR v2 | 41.3 | — |
Category Winners
coding
74.5
math
95.5
reasoning
90.0
general
46.2
multimodal
77.9
agent
75.1