foriequal0/git-trim | ProgramBench

Automatically trims your branches whose tracking remote refs are merged or stray

509

Generated Behavioral Tests

96.5%

Best Score

Muse Wonder Full

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	Muse Wonder Full	96.5%	TBA	577	trace →
2	GPT 6 Astra (max)	96.3%	$9.92	56	trace →
3	Claude Opus 5.5 (xhigh)	94.1%	$23.13	329	trace →
4	Claude Fable 5.1 (xhigh)	90.4%	$30.98	281	trace →
5	Muse Wonder (xhigh)	89.2%	$37.69	463	trace →
6	Claude Opus 5 (xhigh)	86.8%	$34.76	260	trace →
7	Muse Spark 1.3 (max)	85.7%	$3.63	276	trace →
8	Muse Spark 1.3 (xhigh)	79.6%	$1.66	293	trace →
9	GPT 5.5 (xhigh)	79.0%	$7.04	59	trace →
10	GPT-5.6 Sol (xhigh)	78.6%	$5.11	26	trace →
11	Claude Opus 4.8 (xhigh)	76.6%	$25.96	184	trace →
12	GLM-5.2	75.2%	$25.78	245	trace →
13	Gemini 3.7 Flash	71.7%	$1.49	102	trace →
14	GPT 5.5 (high)	65.0%	$2.95	41	trace →
15	Claude Opus 4.6	64.6%	$11.85	321	trace →
16	Muse Spark 1.1 (xhigh)	61.5%	$0.68	225	trace →
17	Muse Spark 1.2 (xhigh)	60.5%	$0.59	166	trace →
18	Gemini 3.6 Flash	58.2%	$5.85	151	trace →
19	Claude Sonnet 4.6	57.0%	$13.30	334	trace →
20	Gemini 3.5 Flash	53.4%	$4.22	141	trace →
21	Claude Opus 4.7	52.7%	$2.67	82	trace →
22	GPT 5.5	50.7%	$0.91	16	trace →
23	Claude Opus 4.7 (xhigh)	48.5%	$4.29	114	trace →
24	GPT-5.6 Sol	47.3%	$0.44	8	trace →
25	GPT 5.4	45.4%	$0.28	15	trace →
26	Claude Haiku 4.5	41.1%	$0.70	138	trace →
27	Gemini 3 Flash	38.9%	$0.37	75	trace →
28	GPT 5.4 mini	11.6%	$0.06	13	trace →
29	GPT 5 mini	10.4%	$0.02	17	trace →
30	Gemini 3.1 Pro	n/a	$1.57	98	trace →

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