DuelLabGameBench 2
Menu

Model results · Alibaba

Qwen3.8 Omni Flash

No Overall rank: results for 2 of 3 reasoning settings. Intensive not supported.

Score · relative 0–100
53.8
Rank
—
Status
Provisional
Players generated
2026-11-07
Match evidence updated
2026-11-10
Compare in charts View player programs Back to leaderboard

Reasoning settings compared

  1. Intensive

    Not supported by this model.

  2. Reasoning enabledBalanced

    69.7 · no rank on its setting leaderboard

    Cost per program
    $0.0376 · provider-reported
    Working programs
    7 of 7
    API time
    15.0 min (observed mean, descriptive)

    31.7 points above Reasoning disabled, at 9.4× the cost

    Not compared: no official rank on its setting leaderboard; its score does not cover every game in the cost comparison.

    • No alternatives: this setting is not compared (no official rank on its setting leaderboard).
  3. Reasoning disabledBaseline

    38.0 · rank 54 on the Baseline leaderboard

    Cost per program
    $0.00401 · provider-reported
    Working programs
    8 of 8
    API time
    60 s (observed mean, descriptive)
    • Higher score for $0.00401 or less: GLM-5.3 Flash Low · Baseline: 46.7 for $0.00373 (provider-reported)
    • Cheapest scoring at least as high: Particle-1 None · Baseline: 39.8 for $0.00221 (provider-reported) (one of 5 such settings)

Score and cost of every ranked setting

The chart needs JavaScript. The setting cards hold this model's results, and Charts compares every model.

Each setting is placed by its published score and its generation cost per player program, the mean over the game cells including repairs and failed attempts. Higher is better; further left costs less (logarithmic axis, with recorded $0 in its own lane). Filled marks are compared; hollow marks are not (the reason is in the table). The line marks the best value across ranked reasoning settings: no ranked setting with complete cost evidence scores at least as high for the same or less. It includes models without an Overall rank, so it can differ from the line on Charts. Costs mix provider-reported charges and list-price estimates; differences in score points are not tested for significance.
Show the numbers

The table needs JavaScript.

Available settings and grouping

Available settings

Reasoning disabledReasoning enabled

Checked on 2026-11-07 · Provider documentation

How settings are grouped

  • Baseline: Reasoning disabled — tested and included
  • Balanced: Reasoning enabled — tested and included
  • Intensive: Not available for this model

Scores by game

  • Go49.3
  • Chess48.0
  • Shogi38.1
  • Xiangqi59.5
  • Hex58.5
  • International Draughts41.4
  • Game of the Amazons39.0
  • Tumbleweed82.8

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: Reasoning enabled (69.7).

Across included settings, the highest game score is Tumbleweed (82.8); the lowest is Shogi (38.1). 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
qwen/qwen3.8-omni-flash
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
131072 (24 of 24 indexed attempts)

Generation evidence

  • Reasoning disabled: 8 contributing player programs across 8 games
  • Reasoning enabled: 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)
GoBalanced1 working of 1 attempt65.61635.8294 / 143 / 127
GoBaseline1 working of 1 attempt33.11314.0105 / 154 / 305
ChessBalanced1 working of 1 attempt53.41496.4125 / 299 / 108
ChessBaseline1 working of 1 attempt42.61411.92 / 431 / 99
ShogiBalanced1 working of 1 attempt44.21377.8149 / 148 / 251
ShogiBaseline1 working of 1 attempt32.01268.297 / 124 / 327
XiangqiBalanced1 working of 1 attempt79.41747.8352 / 143 / 49
XiangqiBaseline1 working of 1 attempt39.61387.743 / 327 / 174
HexBalanced1 working of 1 attempt87.82202.5506 / 0 / 76
HexBaseline1 working of 1 attempt29.21140.4163 / 0 / 419
International DraughtsBaseline1 working of 1 attempt41.41294.8202 / 16 / 336
Game of the AmazonsBalanced1 working of 1 attempt74.61792.9425 / 0 / 153
Game of the AmazonsBaseline1 working of 1 attempt3.4545.515 / 0 / 563
TumbleweedBalanced1 working of 1 attempt83.12097.0478 / 0 / 104
TumbleweedBaseline1 working of 1 attempt82.52080.9474 / 0 / 108

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

0 model-code generation failures; 1 provider/infrastructure failures excluded from the generation denominator. Runtime player faults are separate: 0 recorded one-sided faults across contributing program rows. A crash, an illegal move, or a timeout repeated by a later quiet replay, pinned on one program, also counts as that program's loss in its match record; other faults 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

Reasoning enabled: 71% / 29% / 0% · n 7Reasoning disabled: 63% / 37% / 0% · n 8 No Overall rankIntensive not supported

Generation cost, all settings combined

$0.02average estimated cost per game and reasoning setting

Median
$0.01
Range
$0.0026–$0.08
Provider-reported total
$0.33
Cost data
16 combinations · 16 runs
Price source
Provider-reported
Price list
boardgame-list-prices-2026-10-09
Average output
41k tokens
Output range
3.0k–156k tokens
Runs with output-token data
16
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