Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
coding
| Benchmark | GPT-5.3 Codex | Qwen3 VL 30B A3B Thinking |
|---|---|---|
| SWE-Bench Pro | 56.8 | — |
reasoning
| Benchmark | GPT-5.3 Codex | Qwen3 VL 30B A3B Thinking |
|---|---|---|
| GPQA Diamond | — | 74.4 |
math
| Benchmark | GPT-5.3 Codex | Qwen3 VL 30B A3B Thinking |
|---|---|---|
| AIME 2025 | — | 83.1 |
multimodal
| Benchmark | GPT-5.3 Codex | Qwen3 VL 30B A3B Thinking |
|---|---|---|
| MMMU-Pro | — | 63 |
| CharXiv Reasoning | — | 56.6 |
| ScreenSpot Pro | — | 57.3 |
general
| Benchmark | GPT-5.3 Codex | Qwen3 VL 30B A3B Thinking |
|---|---|---|
| SimpleQA | — | 23.9 |
agent
| Benchmark | GPT-5.3 Codex | Qwen3 VL 30B A3B Thinking |
|---|---|---|
| OSWorld | — | 30.6 |
Category Winners
coding
45.1
math
Q
Qwen3 VL 30B A3B Thinking57.4
reasoning
Q
Qwen3 VL 30B A3B Thinking62.5
general
Q
Qwen3 VL 30B A3B Thinking29.2
multimodal
Q
Qwen3 VL 30B A3B Thinking25.2
agent
Q
Qwen3 VL 30B A3B Thinking22.5