Neura Intelligence Index
Category Comparison
Benchmark-by-Benchmark
reasoning
| Benchmark | Gemini 2.5 Flash-Lite | Llama 3.3 70B Instruct |
|---|---|---|
| GPQA Diamond | 64.6 | 50.5 |
| Humanity's Last Exam | 5.1 | — |
math
| Benchmark | Gemini 2.5 Flash-Lite | Llama 3.3 70B Instruct |
|---|---|---|
| AIME 2025 | 49.8 | — |
coding
| Benchmark | Gemini 2.5 Flash-Lite | Llama 3.3 70B Instruct |
|---|---|---|
| SWE-bench Verified | 31.6 | — |
multimodal
| Benchmark | Gemini 2.5 Flash-Lite | Llama 3.3 70B Instruct |
|---|---|---|
| MMMU | 72.9 | — |
general
| Benchmark | Gemini 2.5 Flash-Lite | Llama 3.3 70B Instruct |
|---|---|---|
| SimpleQA | 10.7 | — |
Category Winners
coding
1.9
math
7.9
reasoning
25.4
general
14.4
multimodal
64.4