DuelLabGameBench 2
Menu

Model results · Tencent

Hy3 Preview

No Overall rank: results for 2 of 3 reasoning settings. Missing Balanced.

Score · relative 0–100
31.4
Rank
—
Status
Provisional
Players generated
2026-07-02 – 2026-07-03
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)32.6
  • None (Baseline group, tested and included)30.3

Every tested setting is included in the overall score.

Available settings and grouping

Available settings

NoneLowHigh

Checked on 2026-07-21 · Provider documentation

How settings are grouped

  • Baseline: None — tested and included
  • Balanced: Low — available, not tested in the current results
  • Intensive: High — tested and included

Scores by game

  • Go71.8
  • Chess38.4
  • Shogi1 of 2 failed16.1
  • Xiangqi46.0
  • Hex1 of 3 failed22.7
  • International Draughts1 of 2 failed7.1
  • Game of the Amazons1 of 2 failed12.8
  • Tumbleweed36.6

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 (32.6).

Across included settings, the highest game score is Go (71.8); the lowest is International Draughts (7.1). 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/hy3-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 25 indexed attempts)

Generation evidence

  • None: 6 contributing player programs across 6 games
  • High: 7 contributing player programs across 7 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 attempt65.21619.4286 / 79 / 151
GoBaseline1 working of 1 attempt78.41775.9355 / 79 / 82
ChessIntensive1 working of 1 attempt39.11387.28 / 344 / 128
ChessBaseline1 working of 1 attempt37.71376.20 / 347 / 133
ShogiBaseline1 working of 1 attempt32.21270.483 / 119 / 292
XiangqiIntensive1 working of 1 attempt43.71426.630 / 359 / 109
XiangqiBaseline1 working of 1 attempt48.21461.281 / 301 / 116
HexIntensive1 working of 2 attempts60.21567.9310 / 0 / 220
HexBaseline1 working of 1 attempt15.3918.076 / 0 / 454
International DraughtsIntensive1 working of 1 attempt14.3859.657 / 6 / 439
Game of the AmazonsIntensive1 working of 1 attempt25.61130.0123 / 0 / 407
TumbleweedIntensive1 working of 1 attempt42.51304.9155 / 114 / 269
TumbleweedBaseline1 working of 1 attempt30.71178.641 / 223 / 274

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

4 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: 56% / 22% / 22% · n 9None: 50% / 25% / 25% · n 8 No Overall rankMissing BalancedMany programs failed

Generation cost

$0.0023average estimated cost per game and reasoning setting

Median
$0.0015
Range
$0.000909–$0.0075
Known cost subtotal (incomplete or unverified charges)
$0.04
Cost data
16 combinations · 17 runs
Price source
Provider-reported
Price list
Not recorded for this older data
Average output
7.7k tokens
Output range
1.4k–28k tokens
Runs with output-token data
17
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