Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
general
| Benchmark | Llama 3.1 405B | Qwen3.5-35B-A3B |
|---|---|---|
| Chatbot Arena Elo | 1244 | — |
| MMLU-Pro | 73.3 | — |
| IFEval | 86.4 | — |
| MMMLU | — | 85.2 |
reasoning
| Benchmark | Llama 3.1 405B | Qwen3.5-35B-A3B |
|---|---|---|
| ARC-Challenge | 96.9 | — |
| BigBench-Hard | 85.4 | — |
| GPQA Diamond | 51.1 | 84.2 |
| Humanity's Last Exam | — | 47.4 |
coding
| Benchmark | Llama 3.1 405B | Qwen3.5-35B-A3B |
|---|---|---|
| HumanEval | 89 | — |
| SWE-bench Verified | — | 69.2 |
language
| Benchmark | Llama 3.1 405B | Qwen3.5-35B-A3B |
|---|---|---|
| HellaSwag | 96 | — |
| WinoGrande | 87.6 | — |
safety
| Benchmark | Llama 3.1 405B | Qwen3.5-35B-A3B |
|---|---|---|
| TruthfulQA | 65.3 | — |
multimodal
| Benchmark | Llama 3.1 405B | Qwen3.5-35B-A3B |
|---|---|---|
| MMMU | — | 81.4 |
| MMMU-Pro | — | 75.1 |
| CharXiv Reasoning | — | 77.5 |
| ScreenSpot Pro | — | 68.6 |
agent
| Benchmark | Llama 3.1 405B | Qwen3.5-35B-A3B |
|---|---|---|
| BrowseComp | — | 61 |
Category Winners
coding
69.3
math
72.0
reasoning
Q
Qwen3.5-35B-A3B81.4
general
67.9
language
87.7
multimodal
Q
Qwen3.5-35B-A3B69.7
safety
59.8
agent
Q
Qwen3.5-35B-A3B