← Back to leaderboard  ·  Show all task instances

sibprogrammer/xq

Command-line XML and HTML beautifier and content extractor

1,109 go medium
792
Generated Behavioral Tests
93.8%
Best Score
Claude Opus 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 Claude Opus 5 (xhigh) 93.8% $38.91 253 trace →
2 Muse Spark 1.3 (max) 93.1% $9.86 431 trace →
3 GPT-5.6 Sol (xhigh) 87.6% $4.12 26 trace →
4 Gemini 3.7 Flash 87.6% $2.19 170 trace →
5 GPT 5.5 (xhigh) 86.5% $10.18 99 trace →
6 GLM-5.2 83.4% $20.91 196 trace →
7 Muse Spark 1.3 (xhigh) 82.2% $3.27 412 trace →
8 GPT-5.6 Sol 81.8% $0.77 15 trace →
9 Muse Spark 1.1 (xhigh) 81.6% $0.08 104 trace →
10 Gemini 3.6 Flash 80.6% $3.48 169 trace →
11 Claude Opus 4.8 (xhigh) 80.5% $17.74 119 trace →
12 Claude Opus 4.7 (xhigh) 79.0% $12.57 214 trace →
13 Muse Spark 1.2 (xhigh) 78.8% $1.27 217 trace →
14 Claude Opus 4.6 75.9% $13.82 356 trace →
15 GPT 5.5 (high) 75.8% $3.52 37 trace →
16 Claude Sonnet 4.6 75.0% $26.84 625 trace →
17 Gemini 3.5 Flash 73.7% $5.92 163 trace →
18 GPT 5.5 72.7% $0.87 16 trace →
19 Gemini 3 Flash 63.5% $0.31 111 trace →
20 Gemini 3.1 Pro 63.1% $2.94 158 trace →
21 GPT 5.4 57.8% $0.66 11 trace →
22 Claude Haiku 4.5 37.8% $0.98 103 trace →
23 GPT 5.4 mini 37.0% $0.06 13 trace →
24 Claude Opus 4.7 2.5% $11.86 227 trace →
25 GPT 5 mini 1.5% $0.03 15 trace →

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