Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | GPT-5.5 | Qwen3.8-27B |
|---|---|---|
| GPQA Diamond | 93.6 | 89.2 |
| Humanity's Last Exam | 52.2 | 30.8 |
| ARC-AGI v2 | 85 | — |
coding
| Benchmark | GPT-5.5 | Qwen3.8-27B |
|---|---|---|
| SWE-Bench Pro | 58.6 | 61.7 |
multimodal
| Benchmark | GPT-5.5 | Qwen3.8-27B |
|---|---|---|
| MMMU-Pro | 83.2 | — |
| CharXiv Reasoning | — | 90.2 |
agent
| Benchmark | GPT-5.5 | Qwen3.8-27B |
|---|---|---|
| BrowseComp | 84.4 | — |
| Toolathlon | 55.6 | — |
| MCP Atlas | 75.3 | — |
math
| Benchmark | GPT-5.5 | Qwen3.8-27B |
|---|---|---|
| FrontierMath | 35.4 | — |
general
| Benchmark | GPT-5.5 | Qwen3.8-27B |
|---|---|---|
| MRCR v2 | 74 | — |
Category Winners
coding
Q
Qwen3.8-27B69.8
math
51.0
reasoning
89.6
general
80.4
multimodal
86.5
agent
67.9