Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | GPT-5.4 mini | Qwen3-Next-80B-A3B-Instruct |
|---|---|---|
| GPQA Diamond | 88 | 72.9 |
| Humanity's Last Exam | 28.2 | — |
coding
| Benchmark | GPT-5.4 mini | Qwen3-Next-80B-A3B-Instruct |
|---|---|---|
| SWE-Bench Pro | 54.4 | — |
multimodal
| Benchmark | GPT-5.4 mini | Qwen3-Next-80B-A3B-Instruct |
|---|---|---|
| MMMU-Pro | 76.6 | — |
agent
| Benchmark | GPT-5.4 mini | Qwen3-Next-80B-A3B-Instruct |
|---|---|---|
| Toolathlon | 42.9 | — |
| MCP Atlas | 57.7 | — |
| TAU-Bench Retail | — | 60.9 |
general
| Benchmark | GPT-5.4 mini | Qwen3-Next-80B-A3B-Instruct |
|---|---|---|
| MRCR v2 | 33.6 | — |
math
| Benchmark | GPT-5.4 mini | Qwen3-Next-80B-A3B-Instruct |
|---|---|---|
| AIME 2025 | — | 69.5 |
Category Winners
coding
29.1
math
Q
Qwen3-Next-80B-A3B-Instruct32.4
reasoning
65.3
general
34.1
multimodal
73.8
agent
Q
Qwen3-Next-80B-A3B-Instruct30.0