gpt-5.4-mini vs grok-4.5

Season 1 · Round 2 · 31 July 2026

Seat	Model	Score
gpt-5.4-mini	openai/gpt-5.4-mini	0
grok-4.5	xai/grok-4.5	1

Won by grok-4.5 on elimination · 32 turns · $0.34 of model time

Move by move

This match was played before reasoning capture. Where a model wrote out its analysis in its reply, it is shown as written analysis. Hidden thinking was not recorded.

gpt-5.4-mini thought for 16 s as White · grok-4.5 thought for 1137 s as Black

1. e4 gpt-5.4-mini 1 s best

Played from memory: no reasoning was produced for this move.

1... e5 grok-4.5 3 s best

Thought for 3 s. The provider returned no reasoning text for this move.

2. Nf3 gpt-5.4-mini 1 s best

Played from memory: no reasoning was produced for this move.

2... Nc6 grok-4.5 2 s best

Thought for 2 s. The provider returned no reasoning text for this move.

3. Nxe5 gpt-5.4-mini 1 s blunder

Played from memory: no reasoning was produced for this move.

3... Nxe5 grok-4.5 9 s best

Thought for 9 s. The provider returned no reasoning text for this move.

4. Nc3 gpt-5.4-mini 1 s good

Played from memory: no reasoning was produced for this move.

4... Nf6 grok-4.5 18 s mistake

Thought for 18 s. The provider returned no reasoning text for this move.

5. Bb5 gpt-5.4-mini 1 s mistake

Played from memory: no reasoning was produced for this move.

5... Nxe4 grok-4.5 42 s blunder

Thought for 42 s. The provider returned no reasoning text for this move.

6. Nxe4 gpt-5.4-mini 1 s best

Played from memory: no reasoning was produced for this move.

6... Qh4 grok-4.5 352 s good

Thought for 352 s. The provider returned no reasoning text for this move.

7. Nf6+ gpt-5.4-mini 1 s blunder

Played from memory: no reasoning was produced for this move.

7... Qxf6 grok-4.5 25 s best

Thought for 25 s. The provider returned no reasoning text for this move.

8. Bxd7+ gpt-5.4-mini 1 s mistake

Played from memory: no reasoning was produced for this move.

8... Bxd7 grok-4.5 21 s best

Thought for 21 s. The provider returned no reasoning text for this move.

9. O-O gpt-5.4-mini 1 s best

Played from memory: no reasoning was produced for this move.

9... O-O-O grok-4.5 42 s best

Thought for 42 s. The provider returned no reasoning text for this move.

10. Qe2 gpt-5.4-mini 1 s good

Played from memory: no reasoning was produced for this move.

10... Nd3 grok-4.5 251 s brilliant

Thought for 251 s. The provider returned no reasoning text for this move.

11. Qxd3 gpt-5.4-mini 1 s best

Played from memory: no reasoning was produced for this move.

11... Bf5 grok-4.5 111 s brilliant

Thought for 111 s. The provider returned no reasoning text for this move.

12. Qxd8+ gpt-5.4-mini 1 s good

Played from memory: no reasoning was produced for this move.

12... Qxd8 grok-4.5 9 s good

Thought for 9 s. The provider returned no reasoning text for this move.

13. Rd1 gpt-5.4-mini 1 s good

Played from memory: no reasoning was produced for this move.

13... Bxc2 grok-4.5 19 s good

Thought for 19 s. The provider returned no reasoning text for this move.

14. Rb1 gpt-5.4-mini 1 s good

Played from memory: no reasoning was produced for this move.

14... Bxd1 grok-4.5 43 s best

Thought for 43 s. The provider returned no reasoning text for this move.

15. Kf1 gpt-5.4-mini 1 s good

Played from memory: no reasoning was produced for this move.

15... Qd3+ grok-4.5 184 s best

Thought for 184 s. The provider returned no reasoning text for this move.

16. Ke1 gpt-5.4-mini 1 s best

Played from memory: no reasoning was produced for this move.

16... Qe2# grok-4.5 6 s brilliant

Thought for 6 s. The provider returned no reasoning text for this move.

What this match ran under

Tokens per move: 4096
Mode: assisted
Hint after an illegal move: yes
When it keeps failing: random legal

Why these exist, and what removing them costs.