DuelLabGameBench 2
Menu

Model results · StepFun

Step 3.7 Flash

Rank 62 of 66 in Overall Models. All three benchmark settings tested.

Score · relative 0–100
24.2
Rank
62
Status
Complete 3/3
Rank change
0
Players generated
2026-07-02 – 2026-10-09
Match evidence updated
2026-10-24
Results published
Compare in charts View player programs Back to leaderboard

Scores by reasoning setting

  • High (Intensive group, tested and included)15.5
  • Medium (Balanced group, tested and included)26.4
  • Low (Baseline group, tested and included)30.6

Every tested setting is included in the overall score.

Available settings and grouping

Available settings

LowMediumHigh

Checked on 2026-07-21 · Provider documentation

How settings are grouped

  • Baseline: Low — tested and included
  • Balanced: Medium — tested and included
  • Intensive: High — tested and included

Scores by game

  • Go17.7
  • Chess2 of 3 failed14.5
  • Shogi1 of 3 failed24.1
  • Xiangqi2 of 3 failed13.4
  • Hex1 of 3 failed30.0
  • International Draughts42.6
  • Game of the Amazons35.1
  • Tumbleweed1 of 3 failed16.1

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: Low (30.6).

Across included settings, the highest game score is International Draughts (42.6); the lowest is Xiangqi (13.4). 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
stepfun/step-3.7-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
230400 (26 of 46 indexed attempts)

Generation evidence

  • Low: 6 contributing player programs across 6 games
  • Medium: 6 contributing player programs across 6 games
  • High: 5 contributing player programs across 5 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 attempt9.7976.332 / 27 / 457
GoBalanced1 working of 1 attempt10.4990.014 / 69 / 433
GoBaseline1 working of 1 attempt33.01303.3113 / 96 / 307
ChessBaseline1 working of 1 attempt43.51421.910 / 382 / 88
ShogiBalanced1 working of 1 attempt34.81294.986 / 138 / 270
ShogiBaseline1 working of 1 attempt37.41317.793 / 148 / 253
XiangqiBaseline1 working of 1 attempt40.21398.732 / 320 / 146
HexIntensive1 working of 1 attempt39.31302.9201 / 0 / 329
HexBalanced1 working of 1 attempt50.71445.8260 / 0 / 270
International DraughtsIntensive1 working of 1 attempt33.11163.7140 / 10 / 352
International DraughtsBalanced1 working of 1 attempt45.31348.2199 / 12 / 291
International DraughtsBaseline1 working of 1 attempt49.31408.0210 / 30 / 262
Game of the AmazonsIntensive1 working of 1 attempt25.01121.3120 / 0 / 410
Game of the AmazonsBalanced1 working of 1 attempt39.01310.5193 / 0 / 337
Game of the AmazonsBaseline1 working of 1 attempt41.31338.9205 / 0 / 325
TumbleweedIntensive1 working of 1 attempt17.11013.925 / 118 / 395
TumbleweedBalanced1 working of 1 attempt31.21184.020 / 270 / 248

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

7 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: 0% / 63% / 37% · n 8Medium: 50% / 25% / 25% · n 8Low: 38% / 37% / 25% · n 8 All three benchmark settings testedMany programs needed repairMany programs failed

Generation cost

$0.03average estimated cost per game and reasoning setting

Median
$0.02
Range
$0.0065–$0.11
Known cost subtotal (incomplete or unverified charges)
$0.74
Cost data
24 combinations · 24 runs
Price source
Provider-reported
Price list
boardgame-list-prices-2026-10-09
Average output
24k tokens
Output range
4.4k–93k 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