← Back to leaderboard  ·  Show all task instances

unhappychoice/gittype

A CLI code-typing game that turns your source code into typing challenges

1,075 rs hard
724
Generated Behavioral Tests
81.1%
Best Score
Muse Spark 1.3 (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 Muse Spark 1.3 (xhigh) 81.1% $1.07 217 trace →
2 GPT-5.6 Sol (xhigh) 81.1% $3.38 28 trace →
3 GPT 5.5 (xhigh) 78.5% $11.11 61 trace →
4 Claude Opus 4.8 (xhigh) 73.3% $11.52 136 trace →
5 Muse Spark 1.1 (xhigh) 71.5% $1.21 210 trace →
6 Claude Opus 5 (xhigh) 70.3% $78.23 382 trace →
7 Gemini 3.6 Flash 68.4% $4.57 170 trace →
8 GPT 5.5 (high) 66.3% $3.70 51 trace →
9 GPT-5.6 Sol 65.3% $0.50 10 trace →
10 GLM-5.2 64.9% $12.80 220 trace →
11 Gemini 3.7 Flash 64.5% $2.44 184 trace →
12 GPT 5.5 62.4% $0.74 16 trace →
13 Claude Opus 4.7 60.4% $1.74 71 trace →
14 Claude Opus 4.7 (xhigh) 51.5% $4.73 133 trace →
15 Muse Spark 1.2 (xhigh) 48.5% $1.04 142 trace →
16 GPT 5.4 39.4% $0.25 10 trace →
17 Claude Haiku 4.5 37.0% $0.34 93 trace →
18 Muse Spark 1.3 (max) 36.0% $14.76 635 trace →
19 Gemini 3.5 Flash 35.6% $8.13 207 trace →
20 Gemini 3.1 Pro 32.3% $1.16 65 trace →
21 Gemini 3 Flash 25.7% $0.39 129 trace →
22 GPT 5 mini 14.4% $0.02 17 trace →
23 Claude Sonnet 4.6 3.7% $18.17 396 trace →
24 GPT 5.4 mini 3.7% $0.05 24 trace →
25 Claude Opus 4.6 0.0% $4.14 242 trace →

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