← Back to leaderboard  ·  Show all task instances

svenstaro/miniserve

🌟 For when you really just want to serve some files over HTTP right now!

7,561 rs medium
304
Generated Behavioral Tests
98.0%
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) 98.0% $95.37 426 trace →
2 Claude Opus 4.8 (xhigh) 94.2% $44.46 251 trace →
3 Muse Spark 1.1 (xhigh) 93.1% $0.69 135 trace →
4 GPT 5.5 (xhigh) 92.4% $8.11 57 trace →
5 Gemini 3.7 Flash 88.5% $1.44 120 trace →
6 GPT 5.5 (high) 88.2% $3.94 31 trace →
7 Muse Spark 1.3 (xhigh) 85.9% $0.94 178 trace →
8 GPT-5.6 Sol 85.5% $1.04 15 trace →
9 GPT-5.6 Sol (xhigh) 85.2% $6.80 41 trace →
10 Muse Spark 1.3 (max) 84.5% $8.10 335 trace →
11 Gemini 3.5 Flash 82.6% $5.74 153 trace →
12 GPT 5.5 80.3% $1.24 16 trace →
13 Claude Opus 4.6 78.6% $9.68 137 trace →
14 Muse Spark 1.2 (xhigh) 77.3% $5.44 222 trace →
15 Claude Opus 4.7 (xhigh) 77.0% $2.94 68 trace →
16 Claude Sonnet 4.6 76.0% $25.03 420 trace →
17 GLM-5.2 63.9% $12.95 173 trace →
18 Gemini 3.1 Pro 59.5% $0.76 41 trace →
19 Claude Opus 4.7 59.2% $0.94 30 trace →
20 GPT 5.4 58.6% $0.23 8 trace →
21 Claude Haiku 4.5 45.4% $0.62 107 trace →
22 Gemini 3 Flash 39.8% $0.30 63 trace →
23 GPT 5.4 mini 14.8% $0.05 26 trace →
24 Gemini 3.6 Flash 9.9% $6.99 154 trace →
25 GPT 5 mini 9.5% $0.03 13 trace →

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