Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
general
| Benchmark | Claude Opus 4 | Hy3 |
|---|---|---|
| IFEval | 91.5 | — |
| MMLU-Pro | 83.5 | — |
| Chatbot Arena Elo | 1365 | — |
reasoning
| Benchmark | Claude Opus 4 | Hy3 |
|---|---|---|
| GPQA Diamond | 74.8 | 90.4 |
| BigBench-Hard | 93.2 | — |
multimodal
| Benchmark | Claude Opus 4 | Hy3 |
|---|---|---|
| MMMU | 74.2 | — |
coding
| Benchmark | Claude Opus 4 | Hy3 |
|---|---|---|
| SWE-bench Verified | 55.8 | 78 |
| HumanEval | 95.1 | — |
| SWE-Bench Pro | — | 57.9 |
agent
| Benchmark | Claude Opus 4 | Hy3 |
|---|---|---|
| BrowseComp | — | 84.2 |
| Toolathlon | — | 48.5 |
| MCP Atlas | — | 79.1 |
| APEX Agents | — | 25.6 |
Category Winners
coding
T
Hy365.8
math
75.5
reasoning
T
Hy387.3
general
96.0
multimodal
67.8
agent
T
Hy352.6