Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | GLM-4.7 | MAI-Thinking-1 |
|---|---|---|
| GPQA Diamond | 85.7 | 84.2 |
| Humanity's Last Exam | 42.8 | — |
math
| Benchmark | GLM-4.7 | MAI-Thinking-1 |
|---|---|---|
| AIME 2025 | 95.7 | 97 |
coding
| Benchmark | GLM-4.7 | MAI-Thinking-1 |
|---|---|---|
| SWE-bench Verified | 73.8 | 73.5 |
| TerminalBench | 33.3 | — |
| SWE-Bench Pro | — | 52.8 |
agent
| Benchmark | GLM-4.7 | MAI-Thinking-1 |
|---|---|---|
| BrowseComp | 52 | — |
Category Winners
coding
Z
GLM-4.757.0
math
80.3
reasoning
81.0
agent
Z
GLM-4.726.7