Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
coding
| Benchmark | Gemini 3.5 Flash-Lite | GLM-4.5-Air |
|---|---|---|
| SWE-Bench Pro | 54.2 | — |
| SWE-bench Verified | — | 57.6 |
| TerminalBench | — | 30 |
| SciCode | — | 37.3 |
multimodal
| Benchmark | Gemini 3.5 Flash-Lite | GLM-4.5-Air |
|---|---|---|
| CharXiv Reasoning | 76.5 | — |
general
| Benchmark | Gemini 3.5 Flash-Lite | GLM-4.5-Air |
|---|---|---|
| MRCR v2 | 21.3 | — |
reasoning
| Benchmark | Gemini 3.5 Flash-Lite | GLM-4.5-Air |
|---|---|---|
| GPQA Diamond | — | 75 |
| Humanity's Last Exam | — | 10.6 |
agent
| Benchmark | Gemini 3.5 Flash-Lite | GLM-4.5-Air |
|---|---|---|
| BrowseComp | — | 21.3 |
| TAU-Bench Retail | — | 77.9 |
Category Winners
coding
37.1
reasoning
Z
GLM-4.5-Air37.9
general
15.3
multimodal
54.5
agent
Z
GLM-4.5-Air39.7