← All comparisons
Gemini 2.5 Flash-Lite (Non-reasoning) vs Llama 3.1 Tulu3 405B
Google vs Allen Institute for AI — side-by-side benchmark comparison
| Gemini 2.5 Flash-Lite (Non-reasoning) | Llama 3.1 Tulu3 405B | |
|---|---|---|
| Intelligence Index | 12.7 | 14.1 |
| Coding Index | 7.4 | — |
| Math Index | 35.3 | — |
| Output speed (tok/s) | 294.2 | 0.0 |
| Blended price ($/1M) | $0.17 | $0.00 |
| Time to first token (s) | 0.34s | 0.00s |
| aime | 50.0% | 13.3% |
| aime 25 | 35.3% | — |
| artificial analysis coding index | 7.40 | — |
| artificial analysis intelligence index | 12.70 | 14.10 |
| artificial analysis math index | 35.30 | — |
| gpqa | 47.4% | 51.6% |
| hle | 3.7% | 3.5% |
| ifbench | 31.5% | — |
| lcr | 31.3% | — |
| livecodebench | 40.0% | 29.1% |
| math 500 | 92.6% | 77.8% |
| mmlu pro | 72.4% | 71.6% |
| scicode | 17.7% | 30.2% |
| tau2 | 19.0% | — |
| terminalbench hard | 2.3% | — |
Benchmark data from Artificial Analysis.