Grok 4.20 (Reasoning)
Grok 4.20 (Reasoning) is developed by xAI, released 2026-03-09.
Model ID: grok-4.20-0309-reasoning
Higher rates apply above a context threshold; see the official site
Input price
$1.25 / 1M tok
Output price
$2.50 / 1M tok
Blended price
$1.5625 / 1M tok
Epoch Capabilities Index (ECI)
No public data yet
Epoch Mock AIME accuracy
92.2%
Epoch AI · as of 2026-02-17
Epoch FrontierMath (T1–3) accuracy
44.9%
Epoch AI · as of 2026-02-17
Arena text rating
No public data yet
Arena coding rating (human preference votes, not task accuracy)
No public data yet
Arena math rating
No public data yet
Arena WebDev rating
No public data yet
Official list price from xAI, as recorded by
models.dev.
No third-party resale listings for this model in our latest
reseller snapshot.
See all of xAI's models on its provider page, or browse the full model directory. Full source licenses and sync dates: /data-sources/.