altdesktop/i3-style | ProgramBench

🎨 Make your i3 config a little more stylish.

539

Generated Behavioral Tests

99.3%

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)

25 runs

#	Model	Score	Cost	Calls
1	Claude Opus 5 (xhigh)	99.3%	$39.54	216	trace →
2	Muse Spark 1.3 (max)	99.3%	$5.17	349	trace →
3	Muse Spark 1.3 (xhigh)	97.8%	$2.35	296	trace →
4	GPT-5.6 Sol (xhigh)	93.3%	$4.55	40	trace →
5	GPT 5.5 (xhigh)	93.3%	$6.85	67	trace →
6	Claude Opus 4.7 (xhigh)	92.2%	$18.42	206	trace →
7	GLM-5.2	91.5%	$20.12	197	trace →
8	Claude Opus 4.8 (xhigh)	88.7%	$28.39	141	trace →
9	GPT 5.5 (high)	87.0%	$4.41	43	trace →
10	GPT 5.5	87.0%	$2.21	28	trace →
11	Gemini 3.7 Flash	85.7%	$1.72	131	trace →
12	Muse Spark 1.1 (xhigh)	82.6%	$0.25	181	trace →
13	Claude Opus 4.6	80.0%	$12.87	205	trace →
14	Claude Sonnet 4.6	77.9%	$16.07	322	trace →
15	Muse Spark 1.2 (xhigh)	75.1%	$0.13	138	trace →
16	Claude Opus 4.7	73.5%	$7.30	165	trace →
17	GPT-5.6 Sol	72.2%	$0.60	12	trace →
18	Gemini 3.5 Flash	72.0%	$5.09	139	trace →
19	Gemini 3.6 Flash	69.9%	$3.33	121	trace →
20	GPT 5.4	48.8%	$0.31	8	trace →
21	Gemini 3 Flash	48.8%	$0.53	107	trace →
22	Gemini 3.1 Pro	42.7%	$2.02	133	trace →
23	GPT 5.4 mini	35.4%	$0.02	8	trace →
24	Claude Haiku 4.5	34.9%	$0.85	117	trace →
25	GPT 5 mini	30.2%	$0.02	11	trace →

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