Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
coding
| Benchmark | Llama 3.1 405B | Muse Spark |
|---|---|---|
| HumanEval | 89 | — |
| SWE-bench Verified | — | 77.4 |
| SWE-Bench Pro | — | 52.4 |
language
| Benchmark | Llama 3.1 405B | Muse Spark |
|---|---|---|
| WinoGrande | 87.6 | — |
| HellaSwag | 96 | — |
reasoning
| Benchmark | Llama 3.1 405B | Muse Spark |
|---|---|---|
| BigBench-Hard | 85.4 | — |
| GPQA Diamond | 51.1 | 89.5 |
| ARC-Challenge | 96.9 | — |
| Humanity's Last Exam | — | 58.4 |
| ARC-AGI v2 |
general
| Benchmark | Llama 3.1 405B | Muse Spark |
|---|---|---|
| IFEval | 86.4 | — |
| Chatbot Arena Elo | 1244 | — |
| MMLU-Pro | 73.3 | — |
safety
| Benchmark | Llama 3.1 405B | Muse Spark |
|---|---|---|
| TruthfulQA | 65.3 | — |
multimodal
| Benchmark | Llama 3.1 405B | Muse Spark |
|---|---|---|
| MMMU-Pro | — | 80.4 |
| CharXiv Reasoning | — | 86.4 |
| ScreenSpot Pro | — | 84.1 |
Category Winners
coding
69.3
math
72.0
reasoning
M
Muse Spark75.8
general
67.9
language
87.7
multimodal
M
Muse Spark85.9
safety
59.8