← All comparisons
EXAONE 4.5 33B (Non-reasoning) vs Gemini 2.5 Flash-Lite (Reasoning)
LG AI Research vs Google — side-by-side benchmark comparison
| EXAONE 4.5 33B (Non-reasoning) | Gemini 2.5 Flash-Lite (Reasoning) | |
|---|---|---|
| Intelligence Index | — | 17.6 |
| Coding Index | — | 9.5 |
| Math Index | — | 53.3 |
| Output speed (tok/s) | 0.0 | 275.9 |
| Blended price ($/1M) | $0.00 | $0.17 |
| Time to first token (s) | 0.00s | 17.31s |
| aime | — | 70.3% |
| aime 25 | — | 53.3% |
| artificial analysis coding index | — | 9.50 |
| artificial analysis intelligence index | — | 17.60 |
| artificial analysis math index | — | 53.30 |
| gpqa | — | 62.5% |
| hle | — | 6.4% |
| ifbench | — | 49.9% |
| lcr | — | 51.3% |
| livecodebench | — | 59.3% |
| math 500 | — | 96.9% |
| mmlu pro | — | 75.9% |
| scicode | — | 19.3% |
| tau2 | — | 18.4% |
| terminalbench hard | — | 4.5% |
Benchmark data from Artificial Analysis.