- Score · relative 0–100
- 90.1
- Rank
- 2
- Status
- Complete 3/3
- Rank change
- New
- Players generated
- 2026-11-12
- Match evidence updated
- 2026-11-14
Reasoning settings compared
- HighIntensive
94.9 · rank 3 on the Intensive leaderboard
- Cost per program
- $0.799 · list-price estimate
- Working programs
- 8 of 8
- API time
- 14.3 min (observed mean, descriptive)
5.1 points above Medium
On the best-value line across ranked settings.
- No ranked setting scores higher for $0.799 or less.
- No ranked setting scores at least as high for less.
- MediumBalanced
89.8 · rank 3 on the Balanced leaderboard
- Cost per program
- At least $0.434 (some attempts have no cost record) · list-price estimate
- Working programs
- 8 of 8
- API time
- 8.2 min (observed mean, descriptive)
4.1 points above Low
Not compared: cost is a known subtotal (some attempts have no cost record).
- Higher score for no more than the $0.434 recorded so far: GPT-6.1 Sol XHigh · Intensive: 91.8 for $0.267 (list-price estimate)
- Cheapest scoring at least as high: Pareto 26.10 (no Overall rank) Reasoning controls unavailable · Baseline: 89.9 for $0.0296 (provider-reported) (one of 3 such settings)
- LowBaseline
85.7 · rank 4 on the Baseline leaderboard
- Cost per program
- $0.165 · list-price estimate
- Working programs
- 8 of 8
- API time
- 2.2 min (observed mean, descriptive)
- Higher score for $0.165 or less: Pareto 26.10 (no Overall rank) Reasoning controls unavailable · Baseline: 89.9 for $0.0296 (provider-reported)
- It is also the cheapest of the 4 offered settings scoring at least as high for less.
Score and cost of every ranked setting
The chart needs JavaScript. The setting cards hold this model's results, and Charts compares every model.
Show the numbers
The table needs JavaScript.
Available settings and grouping
Available settings
Checked on 2026-11-12 · Provider documentation
How settings are grouped
- Baseline: Low — tested and included
- Balanced: Medium — tested and included
- Intensive: High — tested and included
Scores by game
What this result shows
Strongest included setting: High (94.9).
Across included settings, the highest game score is Hex (96.3); the lowest is Xiangqi (73.4). 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
- gemini-4-carbon
- Provider route (catalog)
- 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
- 943718 (28 of 28 indexed attempts)
Generation evidence
- Low: 8 contributing player programs across 8 games
- Medium: 8 contributing player programs across 8 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.
| Game | Reasoning group | Programs | Working-program strength | Raw rating | Combined record (W / D / L) |
|---|---|---|---|---|---|
| Go | Intensive | 1 working of 1 attempt | 97.4 | 2293.9 | 580 / 3 / 17 |
| Go | Balanced | 1 working of 1 attempt | 88.3 | 1950.1 | 520 / 0 / 80 |
| Go | Baseline | 1 working of 1 attempt | 96.6 | 2241.6 | 576 / 0 / 24 |
| Chess | Intensive | 1 working of 1 attempt | 96.0 | 2123.3 | 525 / 21 / 16 |
| Chess | Balanced | 1 working of 1 attempt | 91.1 | 1946.7 | 468 / 73 / 21 |
| Chess | Baseline | 1 working of 1 attempt | 86.7 | 1850.2 | 440 / 75 / 47 |
| Shogi | Intensive | 1 working of 1 attempt | 98.0 | 2454.0 | 401 / 3 / 8 |
| Shogi | Balanced | 1 working of 1 attempt | 96.2 | 2296.9 | 547 / 7 / 22 |
| Shogi | Baseline | 1 working of 1 attempt | 84.4 | 1853.9 | 303 / 63 / 52 |
| Xiangqi | Intensive | 1 working of 1 attempt | 77.8 | 1723.3 | 334 / 208 / 36 |
| Xiangqi | Balanced | 1 working of 1 attempt | 79.6 | 1744.9 | 337 / 223 / 18 |
| Xiangqi | Baseline | 1 working of 1 attempt | 62.7 | 1570.7 | 191 / 315 / 72 |
| Hex | Intensive | 1 working of 1 attempt | 98.3 | 2811.3 | 599 / 0 / 13 |
| Hex | Balanced | 1 working of 1 attempt | 93.6 | 2433.4 | 565 / 0 / 47 |
| Hex | Baseline | 1 working of 1 attempt | 96.9 | 2651.7 | 588 / 0 / 24 |
| International Draughts | Intensive | 1 working of 1 attempt | 99.1 | 2747.4 | 575 / 5 / 4 |
| International Draughts | Balanced | 1 working of 1 attempt | 89.0 | 2165.0 | 504 / 17 / 63 |
| International Draughts | Baseline | 1 working of 1 attempt | 80.0 | 1959.7 | 442 / 27 / 115 |
| Game of the Amazons | Intensive | 1 working of 1 attempt | 97.7 | 2658.0 | 596 / 0 / 14 |
| Game of the Amazons | Balanced | 1 working of 1 attempt | 88.5 | 2134.2 | 534 / 0 / 76 |
| Game of the Amazons | Baseline | 1 working of 1 attempt | 93.4 | 2368.5 | 568 / 0 / 42 |
| Tumbleweed | Intensive | 1 working of 1 attempt | 94.8 | 2457.0 | 576 / 0 / 38 |
| Tumbleweed | Balanced | 1 working of 1 attempt | 92.0 | 2341.8 | 558 / 0 / 56 |
| Tumbleweed | Baseline | 1 working of 1 attempt | 84.6 | 2122.6 | 510 / 0 / 104 |
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; 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, 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
High: 88% / 12% / 0% · n 8Medium: 100% / 0% / 0% · n 8Low: 63% / 37% / 0% · n 8 All three benchmark settings testedGeneration cost, all settings combined
at least $0.47average per game and reasoning setting; some attempts have no cost record
- Median
- $0.43
- Range
- $0.08–$1.30
- Known cost subtotal (incomplete or unverified charges)
- $11.18
- Cost data
- 24 combinations · 24 runs
- Price source
- Standard list-price estimate
- Price list
- boardgame-list-prices-2026-11-09
- Average output
- 42k tokens
- Output range
- 6.0k–110k 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