Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | GPT-5.4 nano | Nemotron 4 340B |
|---|---|---|
| GPQA Diamond | 82.8 | — |
| Humanity's Last Exam | 24.3 | — |
| ARC-Challenge | — | 94.8 |
coding
| Benchmark | GPT-5.4 nano | Nemotron 4 340B |
|---|---|---|
| SWE-Bench Pro | 52.4 | — |
| HumanEval | — | 73.2 |
multimodal
| Benchmark | GPT-5.4 nano | Nemotron 4 340B |
|---|---|---|
| MMMU-Pro | 66.1 | — |
agent
| Benchmark | GPT-5.4 nano | Nemotron 4 340B |
|---|---|---|
| Toolathlon | 35.5 | — |
| MCP Atlas | 56.1 | — |
general
| Benchmark | GPT-5.4 nano | Nemotron 4 340B |
|---|---|---|
| MRCR v2 | 33.1 | — |
| Chatbot Arena Elo | — | 1187 |
| MMLU-Pro | — | 62.8 |
language
| Benchmark | GPT-5.4 nano | Nemotron 4 340B |
|---|---|---|
| HellaSwag | — | 93.5 |
| WinoGrande | — | 86.2 |
safety
| Benchmark | GPT-5.4 nano | Nemotron 4 340B |
|---|---|---|
| TruthfulQA | — | 63.8 |
Category Winners
coding
30.2
math
25.7
reasoning
64.2
general
27.9
language
72.7
multimodal
45.0
safety
50.2
agent