Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | GPT-5.2 Pro | Qwen3.7 Max |
|---|---|---|
| GPQA Diamond | 93.2 | 92.4 |
| Humanity's Last Exam | 36.6 | 41.4 |
| ARC-AGI v2 | 54.2 | — |
math
| Benchmark | GPT-5.2 Pro | Qwen3.7 Max |
|---|---|---|
| AIME 2025 | 100 | — |
agent
| Benchmark | GPT-5.2 Pro | Qwen3.7 Max |
|---|---|---|
| BrowseComp | 77.9 | — |
| MCP Atlas | — | 76.4 |
coding
| Benchmark | GPT-5.2 Pro | Qwen3.7 Max |
|---|---|---|
| SWE-bench Verified | — | 80.4 |
| SWE-Bench Pro | — | 60.6 |
| SciCode | — | 53.5 |
general
| Benchmark | GPT-5.2 Pro | Qwen3.7 Max |
|---|---|---|
| MMMLU | — | 90.3 |
Category Winners
coding
Q
Qwen3.7 Max76.3
math
84.0
reasoning
Q
Qwen3.7 Max81.0
general
Q
Qwen3.7 Max76.5
agent
73.5