← 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 Index8.317.0
Coding Index3.114.0
Math Index50.323.7
Output speed (tok/s)0.060.8
Blended price ($/1M)$0.00$0.79
Time to first token (s)0.00s1.30s
aime32.7%
aime 2550.3%23.7%
artificial analysis coding index3.1014.00
artificial analysis intelligence index8.3017.00
artificial analysis math index50.3023.70
gpqa51.5%61.3%
hle5.8%4.7%
ifbench23.0%36.6%
lcr0.0%0.0%
livecodebench51.6%34.3%
math 50090.2%
mmlu pro58.8%76.2%
scicode9.3%29.9%
tau216.4%27.2%
terminalbench hard0.0%6.1%

Benchmark data from Artificial Analysis.