← All comparisons
Molmo2-8B vs Gemini 2.5 Flash-Lite (Non-reasoning)
Allen Institute for AI vs Google — side-by-side benchmark comparison
| Molmo2-8B | Gemini 2.5 Flash-Lite (Non-reasoning) | |
|---|---|---|
| Intelligence Index | 7.3 | 12.7 |
| Coding Index | 4.4 | 7.4 |
| Math Index | — | 35.3 |
| Output speed (tok/s) | 0.0 | 294.2 |
| Blended price ($/1M) | $0.00 | $0.17 |
| Time to first token (s) | 0.00s | 0.34s |
| aime | — | 50.0% |
| aime 25 | — | 35.3% |
| artificial analysis coding index | 4.40 | 7.40 |
| artificial analysis intelligence index | 7.30 | 12.70 |
| artificial analysis math index | — | 35.30 |
| gpqa | 42.5% | 47.4% |
| hle | 4.4% | 3.7% |
| ifbench | 26.9% | 31.5% |
| lcr | 0.0% | 31.3% |
| livecodebench | — | 40.0% |
| math 500 | — | 92.6% |
| mmlu pro | — | 72.4% |
| scicode | 13.3% | 17.7% |
| tau2 | 0.0% | 19.0% |
| terminalbench hard | 0.0% | 2.3% |
Benchmark data from Artificial Analysis.