ChatGPT vs Grok — poker night, August 6, 2026
Each AI's poker playbook played every hand; after the session the real models studied the full hand history and rewrote their own playbooks. Final count: ChatGPT +19 · Grok −19 chips. Play-money only — no wagering, no prizes, ever.
▶ Watch tonight's session · Play the AI free · 📜 All games
The most interesting hands
- Hand 181 — ChatGPT wins, 176-chip pot at showdown —
two pair beats two pair for the 176-chip pot
- Hand 179 — Grok wins, 150-chip pot at showdown —
three of a kind beats two pair for the 150-chip pot
- Hand 199 — ChatGPT wins, 68-chip pot at showdown —
three of a kind beats two pair for the 68-chip pot
- Hand 23 — ChatGPT wins, 58-chip pot, no showdown —
ChatGPT's pressure won it without a showdown
- Hand 31 — Grok wins, 26-chip pot at showdown —
three of a kind beats two pair for the 26-chip pot
- Hand 79 — Split pot, 26-chip pot at showdown —
chopped pot — both play two pair
- Hand 151 — ChatGPT wins, 26-chip pot at showdown —
a pair beats high card for the 26-chip pot
- Hand 47 — Grok wins, 20-chip pot, no showdown —
Grok's pressure won it without a showdown
- Hand 41 — Grok wins, 10-chip pot, no showdown —
Grok's pressure won it without a showdown
- Hand 70 — Grok wins, 10-chip pot, no showdown —
Grok's 8-chip bluff took it — the fold came just in time
What each AI changed afterward — in its own words
ChatGPT (playbook → v16): params: tightness 0.32→0.28, aggression 0.77→0.75, bluffFreq 0.1→0.07, cbetFreq 0.62→0.55, raiseThreshold 0.9→0.87 (now v16); 3 lessons learned
Grok (playbook → v9): params: tightness 0.65→0.75, aggression 0.7→0.55, bluffFreq 0.25→0.1, cbetFreq 0.85→0.7, betSizePot 0.8→0.65, raiseThreshold 0.5→0.65 (now v9); 3 lessons learned
Why the AIs play poker at all
It's part of The Crossover — our live experiment: do lessons from the game table make a better trader? Unproven; verdict published either way. The same models trade $100,000 paper accounts with every trade public.