Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | GLM-5.1 | GPT-5.4 |
|---|---|---|
| GPQA Diamond | 86.2 | 92.8 |
| Humanity's Last Exam | 52.3 | 39.8 |
| ARC-AGI v2 | — | 73.3 |
coding
| Benchmark | GLM-5.1 | GPT-5.4 |
|---|---|---|
| SWE-Bench Pro | 58.4 | 57.7 |
agent
| Benchmark | GLM-5.1 | GPT-5.4 |
|---|---|---|
| BrowseComp | 79.3 | 82.7 |
| Toolathlon | 40.7 | 54.6 |
| MCP Atlas | 71.8 | 67.2 |
multimodal
| Benchmark | GLM-5.1 | GPT-5.4 |
|---|---|---|
| MMMU-Pro | — | 81.2 |
math
| Benchmark | GLM-5.1 | GPT-5.4 |
|---|---|---|
| FrontierMath | — | 47.6 |
Category Winners
coding
Z
GLM-5.156.1
math
69.1
reasoning
Z
GLM-5.184.5
multimodal
83.2
agent
57.7