← Back to leaderboard  ·  Show all task instances

sigoden/argc

A Bash CLI framework, also a Bash command runner.

1,135 rs
995
Generated Behavioral Tests
85.4%
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) 85.4% $119.80 534 trace →
2 Muse Spark 1.3 (max) 72.0% $9.13 465 trace →
3 GPT-5.6 Sol (xhigh) 62.0% $10.74 52 trace →
4 Gemini 3.7 Flash 52.9% $6.28 217 trace →
5 Claude Opus 4.8 (xhigh) 52.2% $48.62 226 trace →
6 Claude Sonnet 4.6 44.1% $79.14 919 trace →
7 GPT 5.5 (xhigh) 42.8% $13.46 111 trace →
8 Claude Opus 4.6 39.8% $30.14 435 trace →
9 Muse Spark 1.1 (xhigh) 39.7% $0.28 204 trace →
10 Gemini 3.6 Flash 36.5% $13.42 232 trace →
11 GPT 5.5 (high) 31.5% $5.08 46 trace →
12 Claude Opus 4.7 (xhigh) 28.0% $6.90 121 trace →
13 GLM-5.2 25.5% $44.32 386 trace →
14 Gemini 3.5 Flash 22.2% $5.78 213 trace →
15 GPT 5.5 18.4% $2.18 29 trace →
16 Claude Haiku 4.5 10.4% $1.47 175 trace →
17 GPT 5.4 10.3% $0.55 16 trace →
18 Gemini 3 Flash 8.8% $0.21 63 trace →
19 Claude Opus 4.7 7.9% $0.61 34 trace →
20 GPT 5 mini 7.3% $0.02 20 trace →
21 Gemini 3.1 Pro 7.1% $1.30 88 trace →
22 GPT-5.6 Sol 0.5% $1.53 25 trace →
23 Muse Spark 1.3 (xhigh) 0.4% $2.89 348 trace →
24 Muse Spark 1.2 (xhigh) 0.4% $3.23 195 trace →
25 GPT 5.4 mini 0.0% $0.02 7 trace →

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