← All comparisons
Gemini 3.5 Flash (minimal) vs Llama 3 Instruct 8B
Google vs Meta — side-by-side benchmark comparison
| Gemini 3.5 Flash (minimal) | Llama 3 Instruct 8B | |
|---|---|---|
| Intelligence Index | 43.3 | 6.4 |
| Coding Index | 47.1 | 4.0 |
| Math Index | — | — |
| Output speed (tok/s) | 216.5 | 83.3 |
| Blended price ($/1M) | $3.38 | $0.07 |
| Time to first token (s) | 0.81s | 0.47s |
| aime | — | 0.0% |
| aime 25 | — | — |
| artificial analysis coding index | 47.10 | 4.00 |
| artificial analysis intelligence index | 43.30 | 6.40 |
| artificial analysis math index | — | — |
| gpqa | 82.8% | 29.6% |
| hle | 23.1% | 5.1% |
| ifbench | 47.3% | 24.6% |
| lcr | 53.3% | 0.0% |
| livecodebench | — | 9.6% |
| math 500 | — | 49.9% |
| mmlu pro | — | 40.5% |
| scicode | 48.8% | 11.9% |
| tau2 | 58.8% | 0.0% |
| terminalbench hard | 46.2% | 0.0% |
Benchmark data from Artificial Analysis.