← All comparisons

Grok 4.20 0309 v2 (Reasoning) vs Grok 4.20 0309 (Non-reasoning)

xAI vs xAI — side-by-side benchmark comparison

Grok 4.20 0309 v2 (Reasoning)Grok 4.20 0309 (Non-reasoning)
Intelligence Index49.329.7
Coding Index40.525.4
Math Index
Output speed (tok/s)216.0202.6
Blended price ($/1M)$3.00$3.00
Time to first token (s)13.32s0.50s
aime
aime 25
artificial analysis coding index40.5025.40
artificial analysis intelligence index49.3029.70
artificial analysis math index
gpqa91.1%78.5%
hle32.2%22.5%
ifbench81.2%47.8%
lcr58.0%18.0%
livecodebench
math 500
mmlu pro
scicode45.6%32.2%
tau293.0%69.6%
terminalbench hard37.9%22.0%

Benchmark data from Artificial Analysis.