Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | Gemini 2.5 Pro Preview 06-05 | Llama 3.3 70B |
|---|---|---|
| GPQA Diamond | 86.4 | 49.5 |
| Humanity's Last Exam | 21.6 | — |
| BigBench-Hard | — | 82.9 |
| ARC-Challenge | — | 95.2 |
coding
| Benchmark | Gemini 2.5 Pro Preview 06-05 | Llama 3.3 70B |
|---|---|---|
| SWE-bench Verified | 67.2 | — |
| HumanEval | — | 88.4 |
multimodal
| Benchmark | Gemini 2.5 Pro Preview 06-05 | Llama 3.3 70B |
|---|---|---|
| MMMU | 82 | — |
general
| Benchmark | Gemini 2.5 Pro Preview 06-05 | Llama 3.3 70B |
|---|---|---|
| SimpleQA | 54 | — |
| MRCR v2 | 16.4 | — |
| MMLU-Pro | — | 68.9 |
| IFEval | — | 85.3 |
| Chatbot Arena Elo |
language
| Benchmark | Gemini 2.5 Pro Preview 06-05 | Llama 3.3 70B |
|---|---|---|
| HellaSwag | — | 94.8 |
| WinoGrande | — | 86.1 |
safety
| Benchmark | Gemini 2.5 Pro Preview 06-05 | Llama 3.3 70B |
|---|---|---|
| TruthfulQA | — | 62.8 |
Category Winners
coding
66.8
math
71.9
reasoning
56.9
general
60.5
language
77.3
multimodal
84.7
safety
43.7