By model and setting

Each model and reasoning setting appears once. If a model produced more than one player program for this game, the row stands for all of them: Programs counts the working attempts out of the counted attempts, Score and Raw rating average the programs (weighted by attempts; the small range is the lowest and highest program rating), and the combined record adds up their matches. Score is relative within this game (0–100). Scores here are working-program strength: failed generation attempts are left out and count as 0 only in a model's headline score.

GB2-05: Hex - Reasoning Variants
Rank Model Reasoning Score Programs Raw rating Combined record (W / D / L)
1Claude Fable 5.2XHigh99.21 working of 1 attempt2985.9607 / 0 / 5
2GPT-6 AstraXHigh98.91 working of 1 attempt2910.6604 / 0 / 8
3Gemini 4 HydrogenXHigh98.81 working of 1 attempt2888.6603 / 0 / 9
4Claude Fable 5.2Low98.31 working of 1 attempt2811.3599 / 0 / 13
5Gemini 4 CarbonHigh98.31 working of 1 attempt2811.3599 / 0 / 13
6Claude Opus 5.5XHigh98.11 working of 1 attempt2777.5597 / 0 / 15
7GPT-6.1 SolXHigh97.81 working of 1 attempt2746.2595 / 0 / 17
8Claude Sonnet 5.5XHigh97.31 working of 1 attempt2689.7591 / 0 / 21
9Gemini 4 CarbonLow96.91 working of 1 attempt2651.7588 / 0 / 24
10GPT-6.1 Terra ProXHigh95.71 working of 1 attempt2554.0579 / 0 / 33
11Gemini 4 ArgonHigh95.51 working of 1 attempt2544.3578 / 0 / 34
12Claude Fable 5.2Medium95.11 working of 1 attempt2516.4575 / 0 / 37
13GPT-6.1 TerraXHigh94.51 working of 1 attempt2481.6571 / 0 / 41
14GPT-6.1 SolMedium94.21 working of 1 attempt2465.0569 / 0 / 43
15Gemini 4 CarbonMedium93.61 working of 1 attempt2433.4565 / 0 / 47
16Pareto 26.10 PreviewDefault93.31 working of 1 attempt2418.3563 / 0 / 49
17GPT-6 AstraLow93.11 working of 1 attempt2410.8562 / 0 / 50
18GLM-5.3 FlashX LegalMax93.11 working of 1 attempt2410.8562 / 0 / 50
19Claude Fable 5.1XHigh92.81 working of 1 attempt2396.3560 / 0 / 52
20GPT-6 SolXHigh92.81 working of 1 attempt2396.3560 / 0 / 52
21Grok 4.7XHigh92.51 working of 1 attempt2382.1558 / 0 / 54
22Grok 4.7Medium92.41 working of 1 attempt2375.1557 / 0 / 55
23Claude Opus 5XHigh92.02 working of 2 attempts2520.1 program range 2081–29591112 / 0 / 112
24Muse WonderXHigh91.81 working of 1 attempt2347.9553 / 0 / 59
25GPT-6 AstraMedium91.61 working of 1 attempt2341.2552 / 0 / 60

Individual player programs

Each row is one accepted player program for GB2-05: Hex; a model can have more than one. Score is the 0–100 result in this game, and its rating and W / D / L record show how it played. Scores are working-program strength: failed generation attempts have no row and count only in a model's headline score.

305 player programs.

ModelReasoningStatusScoreRatingW / D / LMatchesOpponentsNotesGenerated
Claude Fable 5.2E-dcf455a4461ea112XHighRated Interval validated99.22986607 / 0 / 5612 / 612306—2026-11-10First try
Claude Opus 5E-3f3686d02d1503faXHighRated Interval validated99.12959606 / 0 / 6612 / 612306—2026-07-25First try
GPT-6 AstraE-84a392f4c3b1b8dfXHighRated Interval validated98.92911604 / 0 / 8612 / 612306—2026-09-05First try
Gemini 4 HydrogenE-b1c6e00e20ee459fXHighRated Interval validated98.82889603 / 0 / 9612 / 612306—2026-11-12First try
Claude Fable 5.2E-6a05592537637630LowRated Interval validated98.32811599 / 0 / 13612 / 612306—2026-11-09First try
Gemini 4 CarbonE-998e65fb61851c35HighRated Interval validated98.32811599 / 0 / 13612 / 612306—2026-11-12First try
Claude Opus 5.5E-17a0ec0d72fab9d9XHighRated Interval validated98.12778597 / 0 / 15612 / 612306—2026-09-22Repaired
GPT-6.1 SolE-84e2760b1fa4b2a3XHighRated Interval validated97.82746595 / 0 / 17612 / 612306—2026-10-01First try
Claude Sonnet 5.5E-8e760134a0746d5cXHighRated Interval validated97.32690591 / 0 / 21612 / 612306—2026-09-30First try
Gemini 4 CarbonE-60386266e909150dLowRated Interval validated96.92652588 / 0 / 24612 / 612306—2026-11-12Repaired
GPT-6.1 Terra ProE-27fbe994cc9ada39XHighRated Interval validated95.72554579 / 0 / 33612 / 612306—2026-11-01First try
Gemini 4 ArgonE-520d168a984eb89cHighRated Interval validated95.52544578 / 0 / 34612 / 612306—2026-10-21First try
Claude Fable 5.2E-e0c20e2d3622ab2dMediumRated Interval validated95.12516575 / 0 / 37612 / 612306—2026-11-09First try
GPT-6.1 TerraE-3a6293653aaf6437XHighRated Interval validated94.52482571 / 0 / 41612 / 612306—2026-11-01First try
GPT-6.1 SolE-b18d2c85adab44dbMediumRated Interval validated94.22465569 / 0 / 43612 / 612306—2026-10-01First try
Gemini 4 CarbonE-592b7038ca7556d1MediumRated Interval validated93.62433565 / 0 / 47612 / 612306—2026-11-12First try
Pareto 26.10 PreviewE-938c7886323da9c0DefaultRated Interval validated93.32418563 / 0 / 49612 / 612306—2026-10-09First try
GLM-5.3 FlashX LegalE-902f2e2229ddabf0MaxRated Interval validated93.12411562 / 0 / 50612 / 612306—2026-11-03First try
GPT-6 AstraE-bcfaa85bf834db66LowRated Interval validated93.12411562 / 0 / 50612 / 612306—2026-09-05First try
Claude Fable 5.1E-f1eb379197580e01XHighRated Interval validated92.82396560 / 0 / 52612 / 612306—2026-09-08Repaired
GPT-6 SolE-5fda0fd12201228dXHighRated Interval validated92.82396560 / 0 / 52612 / 612306—2026-09-22First try
Grok 4.7E-be483c51fbcfb505XHighRated Interval validated92.52382558 / 0 / 54612 / 612306—2026-09-30First try
Grok 4.7E-f78a8d93b3f3b22dMediumRated Interval validated92.42375557 / 0 / 55612 / 612306—2026-09-30First try
Muse WonderE-3479ecef790ca818XHighRated Interval validated91.82348553 / 0 / 59612 / 612306—2026-10-12First try
GPT-6 AstraE-9bf97597e76765acMediumRated Interval validated91.62341552 / 0 / 60612 / 612306—2026-09-05First try