Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | Nemotron 3 Super (120B A12B) | Qwen3-Next-80B-A3B-Instruct |
|---|---|---|
| GPQA Diamond | 82.7 | 72.9 |
| Humanity's Last Exam | 22.8 | — |
math
| Benchmark | Nemotron 3 Super (120B A12B) | Qwen3-Next-80B-A3B-Instruct |
|---|---|---|
| AIME 2025 | 90.2 | 69.5 |
coding
| Benchmark | Nemotron 3 Super (120B A12B) | Qwen3-Next-80B-A3B-Instruct |
|---|---|---|
| SWE-bench Verified | 53.7 | — |
| TerminalBench | 25.8 | — |
| SciCode | 42.0 | — |
agent
| Benchmark | Nemotron 3 Super (120B A12B) | Qwen3-Next-80B-A3B-Instruct |
|---|---|---|
| BrowseComp | 31.3 | — |
| TAU-Bench Retail | — | 60.9 |
Category Winners
coding
27.2
math
70.3
reasoning
Q
Qwen3-Next-80B-A3B-Instruct59.6
agent
Q
Qwen3-Next-80B-A3B-Instruct30.0