DuelLabGameBench 2
Menu

Model results · Tencent

Hy4 Preview

Rank 29 of 66 in Overall Models. All three benchmark settings tested.

Score · relative 0–100
53.3
Rank
29
Status
Complete 3/3
Rank change
-1
Players generated
2026-09-03
Match evidence updated
2026-10-26
Results published
Compare in charts View player programs Back to leaderboard

Scores by reasoning setting

  • High (Intensive group, tested and included)64.5
  • Low (Balanced group, tested and included)51.1
  • None (Baseline group, tested and included)44.2

Every tested setting is included in the overall score.

Available settings and grouping

Available settings

NoneLowHigh

Checked on 2026-09-03 · Provider documentation

How settings are grouped

  • Baseline: None — tested and included
  • Balanced: Low — tested and included
  • Intensive: High — tested and included

Scores by game

  • Go51.5
  • Chess1 of 3 failed40.6
  • Shogi59.0
  • Xiangqi44.4
  • Hex57.8
  • International Draughts60.3
  • Game of the Amazons60.9
  • Tumbleweed51.6

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: High (64.5).

Across included settings, the highest game score is Game of the Amazons (60.9); the lowest is Chess (40.6). 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
tencent/hy4-preview
Provider route (catalog)
openrouter
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: 8 contributing player programs across 8 games
  • Low: 7 contributing player programs across 7 games
  • High: 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.

GameReasoning groupProgramsWorking-program strengthRaw ratingCombined record (W / D / L)
GoIntensive1 working of 1 attempt55.01517.3151 / 244 / 121
GoBalanced1 working of 1 attempt52.81495.8245 / 33 / 238
GoBaseline1 working of 1 attempt46.51436.1188 / 83 / 245
ChessIntensive1 working of 1 attempt78.71737.3308 / 128 / 44
ChessBaseline1 working of 1 attempt43.21419.47 / 385 / 88
ShogiIntensive1 working of 1 attempt72.51656.6317 / 49 / 128
ShogiBalanced1 working of 1 attempt41.11350.151 / 267 / 176
ShogiBaseline1 working of 1 attempt63.51554.8236 / 118 / 140
XiangqiIntensive1 working of 1 attempt56.41525.2142 / 260 / 96
XiangqiBalanced1 working of 1 attempt35.31360.138 / 261 / 199
XiangqiBaseline1 working of 1 attempt41.61409.936 / 326 / 136
HexIntensive1 working of 1 attempt70.41719.0364 / 0 / 166
HexBalanced1 working of 1 attempt52.41467.3269 / 0 / 261
HexBaseline1 working of 1 attempt50.51443.5259 / 0 / 271
International DraughtsIntensive1 working of 1 attempt81.01964.5386 / 21 / 95
International DraughtsBalanced1 working of 1 attempt60.41586.4278 / 12 / 212
International DraughtsBaseline1 working of 1 attempt39.31257.3170 / 10 / 322
Game of the AmazonsIntensive1 working of 1 attempt67.81684.4351 / 0 / 179
Game of the AmazonsBalanced1 working of 1 attempt77.21838.0405 / 0 / 125
Game of the AmazonsBaseline1 working of 1 attempt37.71293.8186 / 0 / 344
TumbleweedIntensive1 working of 1 attempt34.11214.934 / 271 / 233
TumbleweedBalanced1 working of 1 attempt89.22134.4474 / 0 / 64
TumbleweedBaseline1 working of 1 attempt31.41186.170 / 172 / 296

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: 0 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

High: 88% / 12% / 0% · n 8Low: 75% / 13% / 12% · n 8None: 88% / 12% / 0% · n 8 All three benchmark settings tested

Generation cost

$0.09average estimated cost per game and reasoning setting

Median
$0.11
Range
$0.0094–$0.19
Provider-reported total
$2.18
Cost data
24 combinations · 24 runs
Price source
Provider-reported
Price list
Not recorded for this older data
Average output
34k tokens
Output range
1.3k–70k 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