← Back to leaderboard  ·  Show all task instances

rs/jplot

iTerm2 expvar/JSON monitoring tool

1,178 go easy
583
Generated Behavioral Tests
96.6%
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) 96.6% $2.96 220 trace →
2 Claude Opus 4.8 (xhigh) 94.7% $25.17 129 trace →
3 GPT 5.5 (high) 91.3% $3.76 43 trace →
4 Claude Opus 4.6 89.0% $8.17 139 trace →
5 GPT 5.5 (xhigh) 86.6% $3.95 38 trace →
6 GPT-5.6 Sol (xhigh) 85.6% $2.66 27 trace →
7 Claude Opus 4.7 (xhigh) 83.4% $4.86 95 trace →
8 Muse Spark 1.1 (xhigh) 83.2% $0.57 119 trace →
9 Claude Opus 4.7 82.5% $2.34 85 trace →
10 Muse Spark 1.2 (xhigh) 82.3% $7.40 288 trace →
11 GLM-5.2 78.4% $18.98 177 trace →
12 Gemini 3.7 Flash 77.4% $1.56 135 trace →
13 GPT 5.5 69.0% $1.90 28 trace →
14 Gemini 3.6 Flash 64.8% $6.60 153 trace →
15 Muse Spark 1.3 (xhigh) 57.8% $1.18 198 trace →
16 Claude Opus 5 (xhigh) 50.8% $50.73 235 trace →
17 Claude Sonnet 4.6 47.7% $31.83 344 trace →
18 Gemini 3 Flash 45.1% $0.15 35 trace →
19 Gemini 3.5 Flash 44.9% $3.88 80 trace →
20 GPT 5.4 43.2% $0.12 11 trace →
21 Claude Haiku 4.5 41.9% $0.18 67 trace →
22 GPT 5.4 mini 35.0% $0.02 10 trace →
23 GPT 5 mini 16.3% $0.02 16 trace →
24 Gemini 3.1 Pro 13.7% $1.63 67 trace →
25 GPT-5.6 Sol 2.6% $0.41 11 trace →

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