Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | GPT-5 | Qwen3-Coder 480B A35B Instruct |
|---|---|---|
| GPQA Diamond | 85.7 | — |
| Humanity's Last Exam | 24.8 | — |
math
| Benchmark | GPT-5 | Qwen3-Coder 480B A35B Instruct |
|---|---|---|
| AIME 2025 | 94.6 | — |
| FrontierMath | 26.3 | — |
coding
| Benchmark | GPT-5 | Qwen3-Coder 480B A35B Instruct |
|---|---|---|
| SWE-bench Verified | 74.9 | 69.6 |
multimodal
| Benchmark | GPT-5 | Qwen3-Coder 480B A35B Instruct |
|---|---|---|
| MMMU | 84.2 | — |
| MMMU-Pro | 78.4 | — |
| CharXiv Reasoning | 81.1 | — |
agent
| Benchmark | GPT-5 | Qwen3-Coder 480B A35B Instruct |
|---|---|---|
| BrowseComp | 54.9 | — |
| TAU-Bench Retail | — | 77.5 |
Category Winners
coding
71.2
math
57.3
reasoning
57.9
multimodal
76.6
agent
Q
Qwen3-Coder 480B A35B Instruct77.0