xAI's new flagship Grok: the first version that retrains itself every week from real-world feedback, and it makes up fewer answers than any model tested so far.
Compared with Grok 4.1 · the version it replaces
Retrains itself every week from real-world feedback
Update cadence — how often the model is refreshed
weekly
✦ new in Grok 4.20 · first Grok to do this
Makes up fewer answers than any model tested so far
AA Omniscience — how often the model avoids inventing facts
78%
▲ best of any model tested · no earlier score to compare
Lands among the top-scoring models in public head-to-head votes
Arena Elo — early ranking from people comparing answers blind
~1505-1535
▲ early measurement · nothing earlier to compare
Grok 4.6 is the current model in this line. The card shows how long it has been since the last release against the line's usual gap and its recent ones — the rhythm, not a date.
Newest first. See the full line with every card →
One email a week: what AI models shipped, the actual differences vs the previous version, and what's coming next. No filler.