Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
coding
| Benchmark | Claude 3.5 Sonnet | Gemini 3.1 Flash-Lite |
|---|---|---|
| SWE-bench Verified | 49 | — |
| HumanEval | 92 | — |
language
| Benchmark | Claude 3.5 Sonnet | Gemini 3.1 Flash-Lite |
|---|---|---|
| HellaSwag | 89 | — |
| WinoGrande | 83.2 | — |
general
| Benchmark | Claude 3.5 Sonnet | Gemini 3.1 Flash-Lite |
|---|---|---|
| Chatbot Arena Elo | 1271 | — |
| MMLU-Pro | 78 | — |
| IFEval | 88 | — |
| SimpleQA | — | 43.3 |
| MMMLU |
reasoning
| Benchmark | Claude 3.5 Sonnet | Gemini 3.1 Flash-Lite |
|---|---|---|
| ARC-Challenge | 96.7 | — |
| GPQA Diamond | 65 | 86.9 |
| BigBench-Hard | 89.4 | — |
| Humanity's Last Exam | — | 16 |
multimodal
| Benchmark | Claude 3.5 Sonnet | Gemini 3.1 Flash-Lite |
|---|---|---|
| MMMU | 68.3 | — |
| MMMU-Pro | — | 76.8 |
| CharXiv Reasoning | — | 73.2 |
safety
| Benchmark | Claude 3.5 Sonnet | Gemini 3.1 Flash-Lite |
|---|---|---|
| TruthfulQA | 74.1 | — |
Category Winners
coding
48.4
math
58.6
reasoning
68.3
general
80.5
language
31.6
multimodal
62.1
safety
95.3