← Back to leaderboard  ·  Show all task instances

NikolaDucak/caps-log

A small TUI journaling tool. 📖

370 cpp medium
551
Generated Behavioral Tests
86.2%
Best Score
Muse Spark 1.3 (max)

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 Muse Spark 1.3 (max) 86.2% $8.33 384 trace →
2 Claude Opus 5 (xhigh) 79.7% $58.77 281 trace →
3 GPT-5.6 Sol (xhigh) 79.7% $7.42 56 trace →
4 Muse Spark 1.3 (xhigh) 78.6% $1.40 213 trace →
5 GPT 5.5 (xhigh) 75.0% $8.89 102 trace →
6 GLM-5.2 67.3% $13.14 132 trace →
7 Gemini 3.7 Flash 66.4% $1.28 115 trace →
8 GPT-5.6 Sol 65.2% $1.13 18 trace →
9 GPT 5.5 (high) 65.0% $5.81 53 trace →
10 Gemini 3.6 Flash 64.4% $3.92 139 trace →
11 Claude Opus 4.8 (xhigh) 62.4% $12.66 112 trace →
12 Gemini 3.1 Pro 61.5% $1.70 90 trace →
13 Gemini 3.5 Flash 60.3% $16.34 251 trace →
14 Muse Spark 1.1 (xhigh) 59.9% $0.28 94 trace →
15 Claude Opus 4.6 56.1% $11.12 283 trace →
16 GPT 5.5 54.8% $0.96 17 trace →
17 Muse Spark 1.2 (xhigh) 46.3% $1.11 157 trace →
18 Claude Opus 4.7 42.5% $2.40 70 trace →
19 GPT 5.4 37.7% $0.29 16 trace →
20 Claude Haiku 4.5 34.8% $0.49 97 trace →
21 Gemini 3 Flash 18.9% $0.36 88 trace →
22 GPT 5.4 mini 18.9% $0.02 7 trace →
23 GPT 5 mini 18.9% $0.01 13 trace →
24 Claude Opus 4.7 (xhigh) 10.0% $4.96 84 trace →
25 Claude Sonnet 4.6 10.0% $23.74 425 trace →

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