git-bahn/git-graph | ProgramBench

Command line tool to show clear git graphs arranged for your branching model

568

Generated Behavioral Tests

97.9%

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)	97.9%	$102.01	358	trace →
2	GPT-5.6 Sol (xhigh)	95.6%	$9.75	47	trace →
3	Claude Opus 4.8 (xhigh)	94.7%	$43.65	230	trace →
4	Muse Spark 1.3 (max)	94.0%	$15.28	440	trace →
5	GPT 5.5 (xhigh)	93.7%	$12.72	109	trace →
6	GPT 5.5 (high)	88.0%	$3.35	42	trace →
7	Muse Spark 1.2 (xhigh)	85.2%	$0.57	205	trace →
8	Gemini 3.6 Flash	85.0%	$3.16	114	trace →
9	GLM-5.2	83.3%	$34.67	272	trace →
10	GPT-5.6 Sol	81.2%	$1.14	16	trace →
11	Claude Sonnet 4.6	80.8%	$49.86	658	trace →
12	Gemini 3.5 Flash	79.6%	$10.63	148	trace →
13	Claude Opus 4.7 (xhigh)	78.9%	$4.01	117	trace →
14	GPT 5.5	74.8%	$1.29	21	trace →
15	Claude Haiku 4.5	59.0%	$1.17	166	trace →
16	GPT 5.4	56.2%	$0.20	10	trace →
17	Gemini 3.1 Pro	35.6%	$0.94	49	trace →
18	Muse Spark 1.3 (xhigh)	31.2%	$2.50	300	trace →
19	GPT 5.4 mini	27.6%	$0.07	20	trace →
20	Gemini 3.7 Flash	3.3%	$5.27	204	trace →
21	Claude Opus 4.7	3.3%	$7.75	160	trace →
22	GPT 5 mini	3.3%	$0.02	14	trace →
23	Muse Spark 1.1 (xhigh)	1.2%	$0.43	227	trace →
24	Gemini 3 Flash	1.2%	$0.24	49	trace →
25	Claude Opus 4.6	0.0%	$1.52	216	trace →

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