← Back to leaderboard  ·  Show all task instances

Y2Z/monolith

⬛️ CLI tool and library for saving complete web pages as a single HTML file

15,024 rs medium
712
Generated Behavioral Tests
88.6%
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) 88.6% $5.07 311 trace →
2 Claude Opus 4.8 (xhigh) 85.0% $31.97 190 trace →
3 GLM-5.2 69.4% $15.91 172 trace →
4 Muse Spark 1.3 (xhigh) 69.2% $0.99 171 trace →
5 GPT 5.5 (xhigh) 67.8% $6.59 61 trace →
6 GPT-5.6 Sol (xhigh) 63.9% $5.36 32 trace →
7 Gemini 3.7 Flash 63.1% $1.33 111 trace →
8 Muse Spark 1.1 (xhigh) 61.9% $0.99 201 trace →
9 GPT 5.5 (high) 60.7% $4.38 57 trace →
10 Gemini 3.5 Flash 60.0% $5.46 182 trace →
11 Gemini 3.6 Flash 56.7% $2.89 103 trace →
12 Claude Opus 4.7 (xhigh) 56.3% $6.26 113 trace →
13 GPT-5.6 Sol 55.2% $1.63 21 trace →
14 GPT 5.5 51.8% $1.00 19 trace →
15 Gemini 3 Flash 51.3% $0.33 72 trace →
16 Gemini 3.1 Pro 38.8% $0.95 77 trace →
17 Claude Haiku 4.5 36.8% $0.89 121 trace →
18 GPT 5.4 mini 33.0% $0.03 8 trace →
19 GPT 5.4 25.8% $0.25 8 trace →
20 Muse Spark 1.2 (xhigh) 2.4% $1.22 216 trace →
21 Claude Opus 4.7 2.4% $3.80 118 trace →
22 Claude Sonnet 4.6 2.4% $14.72 406 trace →
23 Claude Opus 4.6 1.3% $12.81 305 trace →
24 GPT 5 mini 1.3% $0.03 16 trace →
25 Claude Opus 5 (xhigh) 0.0% $11.60 127 trace →

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