jesseduffield/lazygit | ProgramBench

simple terminal UI for git commands

855

Generated Behavioral Tests

83.9%

Best Score

Claude Opus 5.5 (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)

30 runs

#	Model	Score	Cost	Calls
1	Claude Opus 5.5 (xhigh)	83.9%	$66.47	449	trace →
2	Muse Wonder (xhigh)	82.8%	$68.52	609	trace →
3	GPT 6 Astra (max)	78.2%	$13.54	64	trace →
4	Claude Fable 5.1 (xhigh)	77.0%	$87.36	472	trace →
5	Claude Opus 5 (xhigh)	66.2%	$47.64	260	trace →
6	GPT 5.5 (high)	63.4%	$3.21	51	trace →
7	GPT 5.5 (xhigh)	62.2%	$7.85	87	trace →
8	Gemini 3.7 Flash	61.3%	$1.09	96	trace →
9	Gemini 3.5 Flash	61.1%	$6.50	202	trace →
10	GPT-5.6 Sol (xhigh)	59.4%	$2.84	20	trace →
11	Gemini 3.6 Flash	58.6%	$2.83	95	trace →
12	Muse Spark 1.3 (max)	56.5%	$6.25	364	trace →
13	Claude Opus 4.6	56.4%	$9.52	188	trace →
14	Muse Spark 1.1 (xhigh)	54.5%	$1.00	145	trace →
15	GLM-5.2	53.1%	$7.45	115	trace →
16	Claude Opus 4.8 (xhigh)	53.0%	$10.12	108	trace →
17	Claude Sonnet 4.6	51.2%	$8.11	206	trace →
18	GPT-5.6 Sol	49.0%	$0.63	11	trace →
19	Muse Spark 1.3 (xhigh)	42.9%	$1.79	276	trace →
20	Muse Spark 1.2 (xhigh)	40.1%	$0.24	66	trace →
21	Gemini 3 Flash	39.4%	$0.16	41	trace →
22	Claude Opus 4.7 (xhigh)	39.3%	$8.39	168	trace →
23	Claude Opus 4.7	36.8%	$0.93	49	trace →
24	Gemini 3.1 Pro	32.3%	$1.29	87	trace →
25	GPT 5.4	28.3%	$0.39	11	trace →
26	Claude Haiku 4.5	26.8%	$0.31	65	trace →
27	GPT 5 mini	24.3%	$0.02	15	trace →
28	GPT 5.4 mini	23.0%	$0.04	118	trace →
29	GPT 5.5	3.3%	$0.92	15	trace →
30	Muse Wonder Full	0.0%	TBA	550	trace →

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