← Back to leaderboard  ·  Show all task instances

axodotdev/oranda

🎁 generate beautiful landing pages for your developer tools

542 rs medium
767
Generated Behavioral Tests
66.1%
Best Score
GPT-5.6 Sol (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-5.6 Sol (xhigh) 66.1% $5.56 37 trace →
2 GPT 5.5 (xhigh) 65.2% $8.72 70 trace →
3 Gemini 3.7 Flash 60.8% $1.74 149 trace →
4 Gemini 3.5 Flash 60.6% $3.92 150 trace →
5 Muse Spark 1.3 (max) 59.8% $7.34 417 trace →
6 GPT 5.5 (high) 59.7% $3.58 39 trace →
7 GLM-5.2 59.3% $17.04 192 trace →
8 GPT-5.6 Sol 58.9% $1.00 20 trace →
9 GPT 5.5 55.3% $1.15 23 trace →
10 Muse Spark 1.2 (xhigh) 54.9% $0.21 112 trace →
11 Claude Opus 4.7 53.6% $1.36 50 trace →
12 Claude Sonnet 4.6 53.6% $21.15 501 trace →
13 Claude Opus 5 (xhigh) 46.2% $60.82 333 trace →
14 Muse Spark 1.3 (xhigh) 44.3% $1.71 281 trace →
15 Gemini 3.1 Pro 44.1% $0.90 80 trace →
16 Claude Opus 4.8 (xhigh) 43.5% $14.98 127 trace →
17 GPT 5.4 42.5% $0.22 10 trace →
18 Gemini 3 Flash 41.6% $0.42 103 trace →
19 Claude Opus 4.6 40.3% $9.62 238 trace →
20 Muse Spark 1.1 (xhigh) 40.0% $0.45 257 trace →
21 Claude Opus 4.7 (xhigh) 38.6% $9.62 153 trace →
22 GPT 5 mini 31.0% $0.03 19 trace →
23 GPT 5.4 mini 19.2% $0.03 9 trace →
24 Claude Haiku 4.5 17.2% $0.29 65 trace →
25 Gemini 3.6 Flash 7.3% $3.88 148 trace →

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