← Back to leaderboard  ·  Show all task instances

guumaster/hostctl

Your dev tool to manage /etc/hosts like a pro!

1,216 go medium
1,051
Generated Behavioral Tests
96.7%
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) 96.7% $32.35 259 trace →
2 Muse Spark 1.3 (max) 95.5% $4.83 242 trace →
3 Gemini 3.7 Flash 93.4% $3.60 214 trace →
4 Claude Opus 4.8 (xhigh) 93.0% $18.99 163 trace →
5 Muse Spark 1.3 (xhigh) 90.0% $1.88 277 trace →
6 Gemini 3.6 Flash 88.6% $4.84 220 trace →
7 GLM-5.2 87.4% $23.52 276 trace →
8 Claude Sonnet 4.6 82.8% $15.89 416 trace →
9 GPT-5.6 Sol (xhigh) 72.7% $5.03 43 trace →
10 GPT 5.5 (xhigh) 71.1% $7.05 66 trace →
11 Claude Opus 4.7 (xhigh) 69.5% $15.28 176 trace →
12 Claude Opus 4.6 69.4% $10.77 253 trace →
13 GPT 5.5 (high) 68.0% $2.92 30 trace →
14 Muse Spark 1.2 (xhigh) 65.7% $1.44 205 trace →
15 Claude Opus 4.7 64.1% $11.77 211 trace →
16 Gemini 3.5 Flash 62.8% $4.16 135 trace →
17 GPT 5.5 62.4% $1.19 17 trace →
18 GPT-5.6 Sol 61.2% $0.84 20 trace →
19 Gemini 3 Flash 53.0% $0.25 62 trace →
20 Gemini 3.1 Pro 40.3% $1.11 70 trace →
21 Claude Haiku 4.5 31.6% $0.85 132 trace →
22 GPT 5.4 28.5% $0.66 17 trace →
23 GPT 5.4 mini 28.4% $0.05 24 trace →
24 GPT 5 mini 24.6% $0.03 24 trace →
25 Muse Spark 1.1 (xhigh) 0.0% $0.16 72 trace →

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