← Back to leaderboard  ·  Show all task instances

sqlite/sqlite

Official Git mirror of the SQLite source tree

9,434 c
13,510
Generated Behavioral Tests
73.3%
Best Score
Muse Spark 1.3 (max)

Results by Model

Hover for details · Click a model to open it · Drag to zoom the x-axis · Shift-drag to pan · The line marks the Pareto frontier (best score per cost)

Per-Model Results

25 runs
# Model Score Cost Calls
1 Muse Spark 1.3 (max) 73.3% $3.14 185 trace →
2 GPT 5.5 (xhigh) 71.6% $14.62 155 trace →
3 Gemini 3.7 Flash 70.8% $0.70 92 trace →
4 GPT 5.5 (high) 68.4% $3.06 45 trace →
5 GLM-5.2 68.1% $15.34 215 trace →
6 Gemini 3.6 Flash 67.9% $3.58 156 trace →
7 Muse Spark 1.1 (xhigh) 67.6% $0.84 169 trace →
8 Muse Spark 1.3 (xhigh) 67.0% $1.20 212 trace →
9 Claude Opus 4.7 67.0% $1.08 37 trace →
10 Muse Spark 1.2 (xhigh) 65.9% $0.53 70 trace →
11 GPT-5.6 Sol 65.8% $0.87 12 trace →
12 Claude Opus 4.7 (xhigh) 64.5% $2.48 68 trace →
13 GPT 5.5 63.7% $1.69 30 trace →
14 Claude Sonnet 4.6 1.3% $28.86 573 trace →
15 GPT-5.6 Sol (xhigh) 1.3% $8.47 46 trace →
16 Gemini 3.5 Flash 1.0% $6.54 277 trace →
17 Gemini 3 Flash 0.9% $0.19 74 trace →
18 Claude Opus 4.8 (xhigh) 0.8% $21.76 135 trace →
19 GPT 5.4 0.6% $0.20 9 trace →
20 GPT 5.4 mini 0.5% $0.02 11 trace →
21 Gemini 3.1 Pro 0.4% $0.85 77 trace →
22 Claude Haiku 4.5 0.3% $0.57 137 trace →
23 GPT 5 mini 0.2% $0.01 14 trace →
24 Claude Opus 5 (xhigh) 0.0% $7.13 70 trace →
25 Claude Opus 4.6 0.0% $4.54 274 trace →

Click a row to replay how that model rebuilt this program, or the model name to open its full run