Scores by reasoning setting

XHigh (Intensive group, tested and included)54.8
Medium (Balanced group, tested and included)51.0
None (Baseline group, tested and included)35.9

Every tested setting is included in the overall score.

Available settings and grouping

Available settings

NoneLowMediumHighXHighMax

Checked on 2026-07-21 · Provider documentation

How settings are grouped

Baseline: None — tested and included
Balanced: Medium — tested and included
Intensive: XHigh — tested and included

Scores by game

Go68.0
Chess50.0
Shogi34.8
Xiangqi43.5
Hex1 of 3 failed38.0
International Draughts53.4
Game of the Amazons44.9
Tumbleweed45.2

Each game counts every generation attempt, a failed one as 0 (“1 of 2 failed” marks games with failed attempts). Scores are relative within each game, across the included settings; they do not compare game difficulty.

What this result shows

Strongest included setting: XHigh (54.8).

Across included settings, the highest game score is Go (68.0); the lowest is Shogi (34.8). Each game counts every generation attempt, a failed one as 0. These are relative scores within each game, not direct comparisons of difficulty.

Match evidence measures these retained programs. It does not establish how a fresh generation will perform. Independent-generation variation is not estimated in this release.

Test protocol and evidence limits

Model identifier: gpt-5.6-terra
Provider route (catalog): openai
Prompt identity / profile: Not retained in this result set / Not retained in this result set
Prompt version: Not retained in this result set
Generation policy: Generation-quality admission policy (see methodology)
Requested output-token limits: Not retained (0 of 28 indexed attempts)

Generation evidence

None: 7 contributing player programs across 7 games
Medium: 8 contributing player programs across 8 games
XHigh: 8 contributing player programs across 8 games

Counts describe contributing programs, not independent replications: repairs and resends are not new independent samples. Independent-generation lineage is not retained.

Game	Reasoning group	Programs	Working-program strength	Raw rating	Combined record (W / D / L)
Go	Intensive	1 working of 1 attempt	72.1	1696.3	326 / 71 / 119
Go	Balanced	1 working of 1 attempt	77.3	1760.5	387 / 3 / 126
Go	Baseline	1 working of 1 attempt	54.6	1512.6	186 / 169 / 161
Chess	Intensive	1 working of 1 attempt	59.7	1548.5	145 / 266 / 69
Chess	Balanced	1 working of 1 attempt	46.2	1442.2	22 / 383 / 75
Chess	Baseline	1 working of 1 attempt	44.3	1427.6	10 / 389 / 81
Shogi	Intensive	1 working of 1 attempt	36.8	1312.0	88 / 152 / 254
Shogi	Balanced	1 working of 1 attempt	34.4	1291.0	40 / 226 / 229
Shogi	Baseline	1 working of 1 attempt	33.2	1280.3	63 / 169 / 262
Xiangqi	Intensive	1 working of 1 attempt	42.3	1415.5	40 / 325 / 133
Xiangqi	Balanced	1 working of 1 attempt	43.8	1427.4	84 / 252 / 162
Xiangqi	Baseline	1 working of 1 attempt	44.5	1432.1	49 / 328 / 121
Hex	Intensive	1 working of 1 attempt	55.6	1508.3	286 / 0 / 244
Hex	Balanced	1 working of 1 attempt	58.3	1542.7	300 / 0 / 230
International Draughts	Intensive	1 working of 1 attempt	80.7	1959.0	380 / 30 / 92
International Draughts	Balanced	1 working of 1 attempt	41.9	1295.4	168 / 39 / 295
International Draughts	Baseline	1 working of 1 attempt	37.5	1229.6	165 / 2 / 335
Game of the Amazons	Intensive	1 working of 1 attempt	44.3	1376.2	221 / 0 / 309
Game of the Amazons	Balanced	1 working of 1 attempt	60.9	1587.6	312 / 0 / 218
Game of the Amazons	Baseline	1 working of 1 attempt	29.6	1188.1	144 / 0 / 386
Tumbleweed	Intensive	1 working of 1 attempt	46.6	1350.6	162 / 142 / 234
Tumbleweed	Balanced	1 working of 1 attempt	45.0	1333.1	159 / 132 / 247
Tumbleweed	Baseline	1 working of 1 attempt	43.9	1320.0	135 / 168 / 235

Each row covers all the model's programs in one game and setting. Strength and raw rating are attempt-weighted means; under them, each program's score and the lowest and highest rating describe the programs (not a confidence interval). The combined record adds up their matches: evidence volume, not strength.

Failures and budgets

1 model-code generation failures; 0 provider/infrastructure failures excluded from the generation denominator. Runtime player faults are separate: 1 recorded one-sided faults across contributing program rows. A crash or illegal move pinned on one program also counts as that program's loss in its match record; other faults, such as timeouts, are left out of rating evidence.

Repair/resend outcomes and available costs are included. Exact retry ceilings and execution budgets are not retained in this result set; see the methodology for the public policy. Do not infer a measured setting from a current provider catalog.

Player program results

XHigh: 75% / 25% / 0% · n 8Medium: 88% / 12% / 0% · n 8None: 88% / 0% / 12% · n 8 All three benchmark settings tested

Generation cost

$0.19average estimated cost per game and reasoning setting

Median: $0.09
Range: $0.06–$0.46
Estimated total for recorded runs: $4.46
Cost data: 24 combinations · 24 runs
Price source: Standard list-price estimate
Price list: boardgame-list-prices-2026-08-08
Average output: 11k tokens
Output range: 2.4k–29k tokens
Runs with output-token data: 24

How cost is measured

Comparison cost gives equal weight to each model, game, and reasoning-setting combination. Recorded generation attempts include repairs and resends when cost evidence is available; runtime compute and service operation are excluded. Cost and output-token counts use separate telemetry denominators. A list-price estimate is not actual expenditure; mixed evidence is not a uniform standard-price comparison.

Test protocol

Combines this model's tested reasoning settings
Player-program generation; retained protocol details are under “Test protocol and evidence limits” above

GPT-5.6 Terra