← Back to leaderboard  ·  Show all task instances

brocode/fblog

Small command-line JSON Log viewer

561 rs medium
978
Generated Behavioral Tests
98.6%
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) 98.6% $3.03 248 trace →
2 Claude Opus 5 (xhigh) 97.9% $70.40 298 trace →
3 GPT-5.6 Sol (xhigh) 96.0% $6.47 43 trace →
4 Claude Opus 4.8 (xhigh) 95.0% $13.69 130 trace →
5 Muse Spark 1.3 (xhigh) 94.6% $1.11 185 trace →
6 Muse Spark 1.2 (xhigh) 93.6% $0.76 273 trace →
7 Gemini 3.7 Flash 90.2% $1.16 107 trace →
8 Gemini 3.5 Flash 89.7% $6.00 236 trace →
9 GPT 5.5 (high) 87.8% $2.71 39 trace →
10 GPT 5.5 (xhigh) 86.2% $5.98 53 trace →
11 Claude Opus 4.6 86.0% $12.29 304 trace →
12 GLM-5.2 84.3% $15.63 195 trace →
13 Gemini 3.6 Flash 80.1% $2.98 138 trace →
14 GPT-5.6 Sol 76.5% $1.45 21 trace →
15 Claude Opus 4.7 74.4% $5.12 127 trace →
16 GPT 5.5 70.0% $0.83 12 trace →
17 Gemini 3.1 Pro 66.7% $1.92 122 trace →
18 Claude Haiku 4.5 38.9% $0.96 144 trace →
19 GPT 5 mini 28.9% $0.02 20 trace →
20 GPT 5.4 mini 10.1% $0.05 20 trace →
21 Claude Opus 4.7 (xhigh) 1.3% $6.80 125 trace →
22 Muse Spark 1.1 (xhigh) 1.3% $0.41 144 trace →
23 Claude Sonnet 4.6 1.3% $13.91 400 trace →
24 GPT 5.4 1.2% $0.24 8 trace →
25 Gemini 3 Flash 1.2% $0.24 77 trace →

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