Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | Beam 1.0 | Claude Opus 4 |
|---|---|---|
| GPQA Diamond | 90.6 | 74.8 |
| Humanity's Last Exam | 36 | — |
| BigBench-Hard | — | 93.2 |
coding
| Benchmark | Beam 1.0 | Claude Opus 4 |
|---|---|---|
| SWE-bench Verified | 81.6 | 55.8 |
| SWE-Bench Pro | 66.1 | — |
| SciCode | 49.2 | — |
| HumanEval | — | 95.1 |
agent
| Benchmark | Beam 1.0 | Claude Opus 4 |
|---|---|---|
| BrowseComp | 76.8 | — |
| MCP Atlas | 80.8 | — |
multimodal
| Benchmark | Beam 1.0 | Claude Opus 4 |
|---|---|---|
| MMMU | — | 74.2 |
general
| Benchmark | Beam 1.0 | Claude Opus 4 |
|---|---|---|
| Chatbot Arena Elo | — | 1365 |
| IFEval | — | 91.5 |
| MMLU-Pro | — | 83.5 |
Category Winners
coding
R
Beam 1.074.8
math
75.5
reasoning
78.9
general
96.0
multimodal
66.8
agent
R
Beam 1.075.2