Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
language
| Benchmark | Phi-3 Medium | Qwen2 72B Instruct |
|---|---|---|
| WinoGrande | 81.4 | — |
| HellaSwag | 86.8 | — |
general
| Benchmark | Phi-3 Medium | Qwen2 72B Instruct |
|---|---|---|
| Chatbot Arena Elo | 1120 | — |
| MMLU-Pro | 53.5 | — |
reasoning
| Benchmark | Phi-3 Medium | Qwen2 72B Instruct |
|---|---|---|
| ARC-Challenge | 91.6 | — |
| GPQA Diamond | — | 42.4 |
safety
| Benchmark | Phi-3 Medium | Qwen2 72B Instruct |
|---|---|---|
| TruthfulQA | 59.1 | — |
coding
| Benchmark | Phi-3 Medium | Qwen2 72B Instruct |
|---|---|---|
| HumanEval | 62.2 | — |
Category Winners
coding
0.5
math
6.4
reasoning
29.1
general
5.2
language
13.9
safety
22.4