← All comparisons
Gemini 3.1 Flash-Lite vs K-EXAONE (Non-reasoning)
Google vs LG AI Research — side-by-side benchmark comparison
| Gemini 3.1 Flash-Lite | K-EXAONE (Non-reasoning) | |
|---|---|---|
| Intelligence Index | 33.5 | 23.4 |
| Coding Index | 30.1 | 13.5 |
| Math Index | — | 44.0 |
| Output speed (tok/s) | 304.9 | 0.0 |
| Blended price ($/1M) | $0.56 | $0.00 |
| Time to first token (s) | 5.00s | 0.00s |
| aime | — | — |
| aime 25 | — | 44.0% |
| artificial analysis coding index | 30.10 | 13.50 |
| artificial analysis intelligence index | 33.50 | 23.40 |
| artificial analysis math index | — | 44.00 |
| gpqa | 82.2% | 69.5% |
| hle | 16.2% | 5.4% |
| ifbench | 77.2% | 39.6% |
| lcr | 65.3% | 47.0% |
| livecodebench | — | — |
| math 500 | — | — |
| mmlu pro | — | 81.0% |
| scicode | 41.9% | 27.0% |
| tau2 | 31.3% | 59.1% |
| terminalbench hard | 24.2% | 6.8% |
Benchmark data from Artificial Analysis.