← Back to leaderboard  ·  Show all task instances

mibk/dupl

a tool for code clone detection

367 go easy
373
Generated Behavioral Tests
92.8%
Best Score
GPT 5.5 (xhigh)

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 GPT 5.5 (xhigh) 92.8% $6.26 67 trace →
2 GPT-5.6 Sol (xhigh) 92.0% $3.57 40 trace →
3 Claude Opus 4.8 (xhigh) 89.5% $9.35 69 trace →
4 Muse Spark 1.3 (max) 87.9% $5.25 296 trace →
5 Muse Spark 1.3 (xhigh) 87.9% $1.53 213 trace →
6 Claude Opus 4.7 (xhigh) 87.1% $10.07 123 trace →
7 GPT 5.5 (high) 86.9% $2.56 34 trace →
8 Claude Opus 4.7 85.0% $4.93 88 trace →
9 Gemini 3.7 Flash 84.2% $3.98 173 trace →
10 GPT 5.5 84.2% $1.02 15 trace →
11 Gemini 3.5 Flash 79.6% $2.77 105 trace →
12 Claude Opus 4.6 79.6% $13.85 215 trace →
13 Muse Spark 1.2 (xhigh) 79.4% $0.89 137 trace →
14 GPT-5.6 Sol 76.1% $0.95 17 trace →
15 Claude Sonnet 4.6 71.6% $45.63 514 trace →
16 Gemini 3.6 Flash 70.8% $2.41 123 trace →
17 Muse Spark 1.1 (xhigh) 69.4% $1.42 290 trace →
18 Gemini 3 Flash 66.5% $0.26 85 trace →
19 GPT 5.4 58.4% $0.45 13 trace →
20 GPT 5.4 mini 57.1% $0.03 9 trace →
21 Claude Haiku 4.5 55.5% $0.69 111 trace →
22 GLM-5.2 53.6% $23.05 166 trace →
23 Gemini 3.1 Pro 48.0% $1.30 71 trace →
24 GPT 5 mini 35.4% $0.02 13 trace →
25 Claude Opus 5 (xhigh) 0.0% $1.91 30 trace →

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