Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | MiniMax M2.1 | Qwen3 Max |
|---|---|---|
| GPQA Diamond | 81 | 62 |
| Humanity's Last Exam | 22 | — |
math
| Benchmark | MiniMax M2.1 | Qwen3 Max |
|---|---|---|
| AIME 2025 | 81 | 81.6 |
coding
| Benchmark | MiniMax M2.1 | Qwen3 Max |
|---|---|---|
| SWE-bench Verified | 67 | 69.6 |
| TerminalBench | 47.9 | — |
| SciCode | 39 | — |
agent
| Benchmark | MiniMax M2.1 | Qwen3 Max |
|---|---|---|
| BrowseComp | 62 | — |
| Toolathlon | 43.5 | — |
Category Winners
coding
Q
Qwen3 Max60.2
math
Q
Qwen3 Max54.7
reasoning
54.0
agent
38.2