Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | Nemotron 3 Ultra (550B A55B) | Qwen3-235B-A22B-Thinking-2507 |
|---|---|---|
| GPQA Diamond | 87 | 81.1 |
| Humanity's Last Exam | 37.4 | 18.2 |
coding
| Benchmark | Nemotron 3 Ultra (550B A55B) | Qwen3-235B-A22B-Thinking-2507 |
|---|---|---|
| SWE-bench Verified | 70.7 | — |
| SciCode | 44.6 | — |
agent
| Benchmark | Nemotron 3 Ultra (550B A55B) | Qwen3-235B-A22B-Thinking-2507 |
|---|---|---|
| BrowseComp | 44.4 | — |
| TAU-Bench Retail | — | 67.8 |
math
| Benchmark | Nemotron 3 Ultra (550B A55B) | Qwen3-235B-A22B-Thinking-2507 |
|---|---|---|
| AIME 2025 | — | 92.3 |
Category Winners
coding
54.7
math
Q
Qwen3-235B-A22B-Thinking-250773.5
reasoning
74.2
agent
Q
Qwen3-235B-A22B-Thinking-250750.1