Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | GLM-4.5 | Nemotron 3 Super (120B A12B) |
|---|---|---|
| GPQA Diamond | 79.1 | 82.7 |
| Humanity's Last Exam | 14.4 | 22.8 |
coding
| Benchmark | GLM-4.5 | Nemotron 3 Super (120B A12B) |
|---|---|---|
| SWE-bench Verified | 64.2 | 53.7 |
| TerminalBench | 37.5 | 25.8 |
| SciCode | 41.7 | 42.0 |
agent
| Benchmark | GLM-4.5 | Nemotron 3 Super (120B A12B) |
|---|---|---|
| BrowseComp | 26.4 | 31.3 |
| TAU-Bench Retail | 79.7 | — |
math
| Benchmark | GLM-4.5 | Nemotron 3 Super (120B A12B) |
|---|---|---|
| AIME 2025 | — | 90.2 |
Category Winners
coding
Z
GLM-4.550.1
math
70.3
reasoning
54.0
agent
Z
GLM-4.542.4