← Back to leaderboard  ·  Show all task instances

sstadick/hck

A sharp cut(1) clone.

738 rs medium
855
Generated Behavioral Tests
99.1%
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) 99.1% $65.89 332 trace →
2 Muse Spark 1.3 (max) 98.9% $2.50 183 trace →
3 Muse Spark 1.3 (xhigh) 98.0% $1.19 194 trace →
4 GPT 5.5 (xhigh) 97.2% $7.85 94 trace →
5 GPT 5.5 (high) 96.0% $2.63 36 trace →
6 Claude Opus 4.8 (xhigh) 96.0% $25.00 151 trace →
7 Claude Opus 4.7 95.7% $5.84 142 trace →
8 GPT-5.6 Sol (xhigh) 95.4% $4.69 34 trace →
9 Gemini 3.7 Flash 94.7% $2.15 130 trace →
10 Claude Opus 4.6 94.6% $11.60 302 trace →
11 Claude Opus 4.7 (xhigh) 94.4% $11.23 119 trace →
12 GPT-5.6 Sol 93.7% $0.62 11 trace →
13 GLM-5.2 93.6% $10.28 136 trace →
14 Muse Spark 1.1 (xhigh) 93.3% $0.83 225 trace →
15 GPT 5.5 91.8% $0.73 13 trace →
16 Gemini 3.5 Flash 91.2% $4.42 177 trace →
17 Gemini 3.6 Flash 91.1% $3.22 153 trace →
18 Claude Sonnet 4.6 87.5% $15.93 374 trace →
19 GPT 5.4 86.8% $0.25 10 trace →
20 GPT 5 mini 80.5% $0.03 13 trace →
21 GPT 5.4 mini 69.5% $0.03 9 trace →
22 Claude Haiku 4.5 64.9% $0.57 107 trace →
23 Gemini 3.1 Pro 58.7% $1.87 129 trace →
24 Gemini 3 Flash 3.3% $0.33 114 trace →
25 Muse Spark 1.2 (xhigh) 0.0% $1.06 161 trace →

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