← Back to leaderboard  ·  Show all task instances

doxygen/doxygen

Official doxygen git repository

6,422 c hard
229
Generated Behavioral Tests
36.7%
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) 36.7% $25.33 205 trace →
2 Gemini 3.7 Flash 36.7% $0.84 84 trace →
3 Claude Opus 4.7 34.5% $1.17 50 trace →
4 Gemini 3.6 Flash 33.6% $2.26 88 trace →
5 Muse Spark 1.1 (xhigh) 31.4% $0.71 232 trace →
6 Claude Opus 4.8 (xhigh) 29.7% $10.90 115 trace →
7 GPT 5.4 29.7% $0.24 7 trace →
8 Muse Spark 1.2 (xhigh) 29.1% $0.43 86 trace →
9 Gemini 3.5 Flash 27.1% $5.23 174 trace →
10 Muse Spark 1.3 (max) 23.1% $6.05 272 trace →
11 GPT-5.6 Sol 22.7% $1.04 12 trace →
12 Claude Opus 4.7 (xhigh) 21.7% $5.38 138 trace →
13 GPT 5.5 21.0% $1.45 20 trace →
14 GPT 5.5 (high) 20.5% $2.41 29 trace →
15 Claude Haiku 4.5 17.8% $0.73 130 trace →
16 GLM-5.2 16.6% $13.16 189 trace →
17 GPT 5 mini 16.6% $0.02 11 trace →
18 Gemini 3.1 Pro 15.7% $1.59 87 trace →
19 GPT 5.4 mini 14.8% $0.02 6 trace →
20 Muse Spark 1.3 (xhigh) 7.8% $1.30 219 trace →
21 GPT 5.5 (xhigh) 1.7% $6.04 51 trace →
22 GPT-5.6 Sol (xhigh) 1.3% $5.20 40 trace →
23 Claude Opus 4.6 1.3% $15.13 328 trace →
24 Claude Sonnet 4.6 1.3% $28.02 525 trace →
25 Gemini 3 Flash 1.3% $0.21 76 trace →

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