Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | GPT-5.6 Luna | Hy3 |
|---|---|---|
| GPQA Diamond | 92.3 | 90.4 |
coding
| Benchmark | GPT-5.6 Luna | Hy3 |
|---|---|---|
| SWE-Bench Pro | 62.7 | 57.9 |
| SWE-bench Verified | — | 78 |
multimodal
| Benchmark | GPT-5.6 Luna | Hy3 |
|---|---|---|
| MMMU-Pro | 78.4 | — |
agent
| Benchmark | GPT-5.6 Luna | Hy3 |
|---|---|---|
| BrowseComp | 83.3 | 84.2 |
| Toolathlon | 53.4 | 48.5 |
| MCP Atlas | — | 79.1 |
| APEX Agents | — | 25.6 |
math
| Benchmark | GPT-5.6 Luna | Hy3 |
|---|---|---|
| FrontierMath | 78.6 | — |
general
| Benchmark | GPT-5.6 Luna | Hy3 |
|---|---|---|
| MRCR v2 | 41.3 | — |
Category Winners
coding
68.1
math
95.5
reasoning
88.8
general
38.0
multimodal
76.9
agent
66.8