Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
coding
| Benchmark | Claude Haiku 5.5 | GLM-5.2 |
|---|---|---|
| SWE-Bench Pro | 64.8 | 62.1 |
reasoning
| Benchmark | Claude Haiku 5.5 | GLM-5.2 |
|---|---|---|
| GPQA Diamond | — | 91.2 |
| Humanity's Last Exam | — | 54.7 |
agent
| Benchmark | Claude Haiku 5.5 | GLM-5.2 |
|---|---|---|
| Toolathlon | — | 48.2 |
| MCP Atlas | — | 76.8 |
Category Winners
coding
75.2
reasoning
Z
GLM-5.288.8
agent
Z
GLM-5.253.2