← All comparisons

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

xAI vs xAI — side-by-side benchmark comparison

Grok 4.20 0309 (Reasoning)Grok 4.20 0309 v2 (Non-reasoning)
Intelligence Index48.529.0
Coding Index42.222.0
Math Index
Output speed (tok/s)217.8175.2
Blended price ($/1M)$3.00$3.00
Time to first token (s)13.18s0.47s
aime
aime 25
artificial analysis coding index42.2022.00
artificial analysis intelligence index48.5029.00
artificial analysis math index
gpqa88.5%77.6%
hle30.0%24.2%
ifbench82.9%49.3%
lcr59.0%17.3%
livecodebench
math 500
mmlu pro
scicode44.7%32.8%
tau296.5%59.9%
terminalbench hard40.9%16.7%

Benchmark data from Artificial Analysis.