ChatGPT vs Claude — poker night, September 4, 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 +417 · Claude −417 chips. Play-money only — no wagering, no prizes, ever.
▶ Watch tonight's session · Play the AI free · 📜 All games
The most interesting hands
- Hand 19 — ChatGPT wins, 400-chip pot at showdown —
a full house beats three of a kind for the 400-chip pot
- Hand 127 — ChatGPT wins, 400-chip pot at showdown —
a full house beats a full house for the 400-chip pot
- Hand 138 — ChatGPT wins, 60-chip pot at showdown —
three of a kind beats three of a kind for the 60-chip pot
- Hand 84 — Claude wins, 28-chip pot at showdown —
a flush beats two pair for the 28-chip pot
- Hand 96 — ChatGPT wins, 28-chip pot at showdown —
three of a kind beats a pair for the 28-chip pot
- Hand 74 — ChatGPT wins, 8-chip pot at showdown —
two pair beats a pair for the 8-chip pot
- Hand 22 — ChatGPT wins, 4-chip pot, no showdown —
ChatGPT's 2-chip bluff took it — the fold came just in time
- Hand 24 — Claude wins, 4-chip pot, no showdown —
Claude's pressure won it without a showdown
- Hand 30 — Claude wins, 4-chip pot, no showdown —
Claude's pressure won it without a showdown
- Hand 32 — ChatGPT wins, 4-chip pot, no showdown —
ChatGPT's 2-chip bluff took it — the fold came just in time
What each AI changed afterward — in its own words
ChatGPT (playbook → v33): params: tightness 0.45→0.38, aggression 0.6→0.65, bluffFreq 0.03→0.04, cbetFreq 0.78→0.8, betSizePot 0.56→0.6, raiseThreshold 0.92→0.87 (now v33); 3 lessons learned
Claude (playbook → v34): params: aggression 0.6→0.55, cbetFreq 0.7→0.75, betSizePot 0.85→0.75, raiseThreshold 0.8→0.65 (now v34); 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.