← Back to leaderboard  ·  Show all task instances

BurntSushi/ripgrep

ripgrep recursively searches directories for a regex pattern while respecting your gitignore

62,855 rs medium
1,992
Generated Behavioral Tests
97.7%
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) 97.7% $126.57 413 trace →
2 Muse Spark 1.3 (max) 92.0% $3.42 266 trace →
3 GPT 5.5 (xhigh) 84.0% $17.56 121 trace →
4 Claude Sonnet 4.6 79.7% $19.39 508 trace →
5 Muse Spark 1.3 (xhigh) 78.5% $1.17 213 trace →
6 Claude Opus 4.8 (xhigh) 78.0% $23.89 158 trace →
7 Muse Spark 1.2 (xhigh) 76.2% $0.38 120 trace →
8 GPT-5.6 Sol (xhigh) 75.9% $7.13 42 trace →
9 GLM-5.2 70.7% $20.64 271 trace →
10 GPT 5.5 (high) 70.5% $3.00 28 trace →
11 GPT-5.6 Sol 67.4% $1.04 15 trace →
12 Gemini 3.6 Flash 67.0% $5.01 184 trace →
13 Gemini 3.7 Flash 67.0% $1.11 151 trace →
14 GPT 5.5 66.0% $1.27 18 trace →
15 Claude Opus 4.7 (xhigh) 62.4% $6.75 168 trace →
16 Gemini 3.5 Flash 50.0% $4.23 146 trace →
17 Claude Opus 4.7 40.7% $2.47 75 trace →
18 GPT 5.4 35.5% $0.26 7 trace →
19 Gemini 3 Flash 35.1% $0.25 77 trace →
20 Gemini 3.1 Pro 29.7% $1.35 103 trace →
21 GPT 5 mini 27.7% $0.03 18 trace →
22 Claude Haiku 4.5 23.7% $0.88 118 trace →
23 Muse Spark 1.1 (xhigh) 2.3% $0.20 138 trace →
24 GPT 5.4 mini 0.2% $0.01 11 trace →
25 Claude Opus 4.6 0.0% $2.74 261 trace →

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