- Score · relative 0–100
- 52.3
- Rank
- 32
- Status
- Complete 3/3
- Rank change
- 0
- Players generated
- 2026-08-21 – 2026-08-22
- Match evidence updated
- 2026-10-26
- Results published
Scores by reasoning setting
Available settings and grouping
Available settings
Checked on 2026-08-21 · Provider documentation
How settings are grouped
- Baseline: Low — tested and included
- Balanced: High — tested and included
- Intensive: Max — tested and included
Scores by game
What this result shows
Strongest included setting: Max (68.3).
Across included settings, the highest game score is Xiangqi (59.5); the lowest is International Draughts (46.5). 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
- stealth/ox-alpha
- 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 (102 of 114 indexed attempts)
Generation evidence
- Low: 28 contributing player programs across 8 games
- High: 26 contributing player programs across 8 games
- Max: 25 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 | 3 working of 3 attempts | 58.6 program scores 66.3 · 66.0 · 43.5 | 1555.1 program range 1407–1631 | 838 / 73 / 637 |
| Go | Balanced | 2 working of 3 attempts | 62.3 program scores 65.6 · 59.0 | 1589.9 program range 1556–1624 | 512 / 218 / 302 |
| Go | Baseline | 3 working of 3 attempts | 42.5 program scores 50.6 · 46.2 · 30.7 | 1395.9 program range 1279–1475 | 444 / 368 / 736 |
| Chess | Intensive | 2 working of 2 attempts | 74.1 program scores 81.4 · 66.7 | 1691.5 program range 1610–1773 | 522 / 352 / 86 |
| Chess | Balanced | 3 working of 3 attempts | 50.2 program scores 54.5 · 48.3 · 47.8 | 1473.2 program range 1454–1507 | 179 / 1037 / 224 |
| Chess | Baseline | 3 working of 3 attempts | 43.1 program scores 44.7 · 42.6 · 41.9 | 1418.0 program range 1409–1431 | 43 / 1106 / 291 |
| Shogi | Intensive | 5 working of 5 attempts | 71.2 program scores 87.3 · 82.0 · 74.8 · 56.3 · 55.7 | 1674.3 program range 1479–1919 | 1612 / 143 / 714 |
| Shogi | Balanced | 7 working of 9 attempts | 57.7 program scores 84.1 · 70.3 · 60.2 · 50.3 · 49.2 · 47.9 · 41.8 | 1515.9 program range 1356–1845 | 1678 / 385 / 1395 |
| Shogi | Baseline | 8 working of 9 attempts | 45.2 program scores 57.7 · 52.5 · 51.8 · 48.3 · 44.1 · 43.0 · 38.0 · 26.5 | 1385.7 program range 1216–1497 | 967 / 1345 / 1640 |
| Xiangqi | Intensive | 3 working of 3 attempts | 71.5 program scores 88.4 · 64.5 · 61.7 | 1682.7 program range 1569–1887 | 765 / 565 / 164 |
| Xiangqi | Balanced | 3 working of 3 attempts | 57.3 program scores 61.0 · 57.1 · 53.8 | 1532.7 program range 1504–1563 | 393 / 873 / 228 |
| Xiangqi | Baseline | 3 working of 3 attempts | 49.6 program scores 56.0 · 47.2 · 45.5 | 1471.8 program range 1440–1522 | 247 / 935 / 312 |
| Hex | Intensive | 3 working of 3 attempts | 62.6 program scores 71.3 · 60.0 · 56.6 | 1606.9 program range 1521–1735 | 969 / 0 / 621 |
| Hex | Balanced | 3 working of 3 attempts | 49.3 program scores 56.0 · 49.9 · 42.0 | 1429.1 program range 1338–1513 | 759 / 0 / 831 |
| Hex | Baseline | 3 working of 3 attempts | 51.7 program scores 57.0 · 53.0 · 45.1 | 1458.8 program range 1377–1525 | 796 / 0 / 794 |
| International Draughts | Intensive | 3 working of 3 attempts | 68.6 program scores 85.4 · 81.1 · 39.4 | 1761.1 program range 1259–2058 | 953 / 80 / 473 |
| International Draughts | Balanced | 2 working of 2 attempts | 28.1 program scores 49.8 · 6.4 | 1041.5 program range 667–1416 | 243 / 21 / 738 |
| International Draughts | Baseline | 3 working of 3 attempts | 42.8 program scores 55.2 · 50.7 · 22.5 | 1311.2 program range 1002–1501 | 572 / 26 / 908 |
| Game of the Amazons | Intensive | 3 working of 3 attempts | 66.7 program scores 75.1 · 64.8 · 60.2 | 1673.4 program range 1578–1801 | 1035 / 0 / 555 |
| Game of the Amazons | Balanced | 3 working of 3 attempts | 45.6 program scores 74.3 · 48.5 · 14.0 | 1381.6 program range 930–1786 | 697 / 0 / 893 |
| Game of the Amazons | Baseline | 3 working of 3 attempts | 39.8 program scores 58.5 · 44.1 · 16.6 | 1304.4 program range 983–1557 | 597 / 0 / 993 |
| Tumbleweed | Intensive | 3 working of 3 attempts | 73.4 program scores 86.5 · 72.7 · 60.8 | 1775.8 program range 1534–2055 | 1101 / 79 / 434 |
| Tumbleweed | Balanced | 3 working of 3 attempts | 61.6 program scores 73.4 · 58.0 · 53.3 | 1558.5 program range 1431–1752 | 797 / 285 / 532 |
| Tumbleweed | Baseline | 2 working of 3 attempts | 30.2 program scores 32.8 · 27.7 | 1172.9 program range 1145–1201 | 120 / 360 / 596 |
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
5 model-code generation failures; 5 provider/infrastructure failures excluded from the generation denominator. Runtime player faults are separate: 176 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
Max: 88% / 12% / 0% · n 25High: 73% / 17% / 10% · n 29Low: 63% / 30% / 7% · n 30 All three benchmark settings testedGeneration cost
$0.00average estimated cost per game and reasoning setting
- Median
- $0.00
- Range
- $0.00–$0.00
- Known cost subtotal (incomplete or unverified charges)
- $0.00
- Cost data
- 24 combinations · 88 runs
- Price source
- Provider-reported
- Price list
- Not recorded for this older data
- Average output
- 23k tokens
- Output range
- 1.7k–85k tokens
- Runs with output-token data
- 88
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