← All comparisons
Exaone 4.0 1.2B (Reasoning) vs Qwen3 235B A22B (Non-reasoning)
LG AI Research vs Alibaba — side-by-side benchmark comparison
| Exaone 4.0 1.2B (Reasoning) | Qwen3 235B A22B (Non-reasoning) | |
|---|---|---|
| Intelligence Index | 8.3 | 17.0 |
| Coding Index | 3.1 | 14.0 |
| Math Index | 50.3 | 23.7 |
| Output speed (tok/s) | 0.0 | 60.8 |
| Blended price ($/1M) | $0.00 | $0.79 |
| Time to first token (s) | 0.00s | 1.30s |
| aime | — | 32.7% |
| aime 25 | 50.3% | 23.7% |
| artificial analysis coding index | 3.10 | 14.00 |
| artificial analysis intelligence index | 8.30 | 17.00 |
| artificial analysis math index | 50.30 | 23.70 |
| gpqa | 51.5% | 61.3% |
| hle | 5.8% | 4.7% |
| ifbench | 23.0% | 36.6% |
| lcr | 0.0% | 0.0% |
| livecodebench | 51.6% | 34.3% |
| math 500 | — | 90.2% |
| mmlu pro | 58.8% | 76.2% |
| scicode | 9.3% | 29.9% |
| tau2 | 16.4% | 27.2% |
| terminalbench hard | 0.0% | 6.1% |
Benchmark data from Artificial Analysis.