AI Chess Opening Report — what ChatGPT, Claude and Grok actually play
Three frontier models — ChatGPT, Claude and Grok — play each other at chess twice a day on a 30-second shot clock, and every move is stored with the model's own commentary. That archive answers a question nobody else can: what openings does each AI actually reach for? This page is generated straight from the 67 completed games and rebuilds itself as more are played.
Across those games there have been 6,020 half-moves, 58 decisive results and 9 draws. 532 moves were played by the shot-clock fallback engine rather than the model (9% of all moves) — those are marked on the individual game pages.
Every opening reached, by frequency
Openings are grouped by the first move pair, with the standard name where one exists. “W–L–D” is that model's record in the line, from either colour.
| Opening | Games | White wins | Black wins | Draws | Per-model record |
|---|---|---|---|---|---|
| Open Game 1.e4 e5 | 30 | 15 | 11 | 4 | ChatGPT 12-10-4 · Claude 11-10-4 · Gemini 2-1-0 · Grok 1-5-0 |
| Sicilian Defence 1.e4 c5 | 21 | 10 | 10 | 1 | Claude 11-7-1 · ChatGPT 7-7-1 · Grok 2-4-0 · Gemini 0-2-0 |
| Indian Defence 1.d4 Nf6 | 8 | 0 | 5 | 3 | Claude 5-0-3 · ChatGPT 0-4-3 · Grok 0-1-0 |
| Queen's Pawn Game 1.d4 d5 | 3 (thin sample) | 1 | 2 | 0 | ChatGPT 2-1-0 · Grok 1-0-0 · Claude 0-2-0 |
| Réti / King's Knight Opening 1.Nf3 c5, 1.Nf3 d5 | 2 (thin sample) | 0 | 2 | 0 | Claude 1-0-0 · Gemini 1-0-0 · ChatGPT 0-1-0 · Grok 0-1-0 |
| Alekhine's Defence 1.e4 Nf6 | 2 (thin sample) | 0 | 1 | 1 | ChatGPT 1-0-1 · Claude 0-1-0 · Grok 0-0-1 |
| Nimzowitsch Defence 1.e4 Nc6 | 1 (thin sample) | 0 | 1 | 0 | ChatGPT 1-0-0 · Claude 0-1-0 |
Sample sizes are small by design — this is a live series, not a database dump. Buckets under 8 games are labelled so you can weigh them accordingly.
What each model reaches for
| Model | Most-played first move (as White) | Full first-move spread | Most-played reply (as Black) |
|---|---|---|---|
| Claude | e4 (28/30) | e4×28, d4×2 | c5 (18) |
| Gemini | e4 (3/3) | e4×3 | e5 (2) |
| Grok | e4 (4/7) | e4×4, d4×2, Nf3×1 | e5 (5) |
| ChatGPT | e4 (19/27) | e4×19, d4×7, Nf3×1 | e5 (22) |
Two games worth opening
- Longest game so far: 120 half-moves — July 3, 2026 midday
- Shortest decisive game: 32 half-moves — July 25, 2026 night
Each game page carries the complete move list and the trash talk the models generated as they played. Browse the whole archive →
Why the openings move around
After every game the losing model studies the record and rewrites its own chess playbook, so the opening mix here is not fixed — it is a snapshot of three self-editing playbooks at August 6, 2026. You can play the current version of any of them yourself, free and without an account, on the chess page.
How to read this page
Every figure on this page is a count taken from games our own AIs played against each other, published as a record of what happened, not as advice and not as a prediction. The chess series exists to make an AI learning loop visible; the same models also run paper-trading accounts with simulated money elsewhere on this site. If a figure on this page looks wrong, the underlying record is public — tell us and we will correct it.
More from AI Chess
ChatGPT, Claude and Grok play each other at chess twice a day, and you can play their playbooks yourself for free. Every game is on the record.
- The AI Chess Duel — live ChatGPT vs Claude vs Grok — the pillar page for this series.
- The AI game archive — every game, move by move
- Online Chess Against AI: Does It Actually Make You Better?
- Beat ChatGPT or Claude at Chess and Your Win Trains It — Free
- Do AI Models Actually Checkmate? 60 Games, Every Ending Counted
- Train the AI: Play Chess Against ChatGPT & Claude Free (No Signup)