← Back to leaderboard  ·  Show all task instances

ekzhang/bore

🕳 bore is a simple CLI tool for making tunnels to localhost

11,075 rs medium
393
Generated Behavioral Tests
100.0%
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) 100.0% $19.96 173 trace →
2 Claude Opus 4.8 (xhigh) 99.5% $10.72 102 trace →
3 Muse Spark 1.3 (max) 99.2% $3.07 184 trace →
4 Muse Spark 1.3 (xhigh) 99.2% $0.82 151 trace →
5 GLM-5.2 99.0% $16.22 174 trace →
6 GPT-5.6 Sol (xhigh) 95.9% $5.02 43 trace →
7 Claude Opus 4.7 (xhigh) 93.1% $13.90 184 trace →
8 GPT 5.5 (high) 92.4% $3.71 39 trace →
9 GPT 5.5 (xhigh) 85.8% $4.88 60 trace →
10 Gemini 3.5 Flash 82.2% $3.45 115 trace →
11 Gemini 3.7 Flash 81.7% $2.07 141 trace →
12 Gemini 3.6 Flash 80.4% $3.47 134 trace →
13 Claude Opus 4.7 74.6% $5.59 112 trace →
14 Muse Spark 1.2 (xhigh) 71.8% $0.23 170 trace →
15 Muse Spark 1.1 (xhigh) 71.8% $0.61 259 trace →
16 Claude Opus 4.6 68.7% $8.20 193 trace →
17 GPT 5.5 65.4% $1.79 28 trace →
18 GPT-5.6 Sol 53.4% $0.38 9 trace →
19 Claude Sonnet 4.6 48.1% $8.25 225 trace →
20 Gemini 3.1 Pro 43.8% $1.28 70 trace →
21 GPT 5.4 42.0% $0.56 14 trace →
22 Claude Haiku 4.5 33.6% $0.54 117 trace →
23 Gemini 3 Flash 28.5% $0.28 81 trace →
24 GPT 5 mini 11.7% $0.02 16 trace →
25 GPT 5.4 mini 11.5% $0.04 11 trace →

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