Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | GPT-4.5 | Qwen3 235B A22B |
|---|---|---|
| GPQA Diamond | 69.5 | 47.5 |
coding
| Benchmark | GPT-4.5 | Qwen3 235B A22B |
|---|---|---|
| SWE-bench Verified | 38 | — |
multimodal
| Benchmark | GPT-4.5 | Qwen3 235B A22B |
|---|---|---|
| MMMU | 75.2 | — |
| CharXiv Reasoning | 55.4 | — |
agent
| Benchmark | GPT-4.5 | Qwen3 235B A22B |
|---|---|---|
| TAU-Bench Retail | 68.4 | — |
math
| Benchmark | GPT-4.5 | Qwen3 235B A22B |
|---|---|---|
| AIME 2025 | — | 81.5 |
Category Winners
coding
4.7
math
Q
Qwen3 235B A22B54.5
reasoning
53.7
general
70.2
multimodal
39.6
agent
51.9