jarun/nnn | ProgramBench

n³ The unorthodox terminal file manager

477

Generated Behavioral Tests

98.7%

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)

29 runs

#	Model	Score	Cost	Calls
1	Muse Wonder Full	98.7%	TBA	697	trace →
2	Claude Opus 5 (xhigh)	98.5%	$78.92	381	trace →
3	Claude Opus 5.5 (xhigh)	98.3%	$58.62	444	trace →
4	GLM-5.2	98.3%	$5.83	114	trace →
5	Muse Spark 1.1 (xhigh)	98.3%	$0.18	62	trace →
6	Muse Spark 1.3 (xhigh)	98.1%	$0.77	152	trace →
7	GPT 5.5 (high)	98.1%	$3.95	36	trace →
8	Claude Opus 4.8 (xhigh)	98.1%	$15.25	151	trace →
9	Claude Opus 4.6	98.1%	$14.07	334	trace →
10	Muse Spark 1.3 (max)	97.9%	$4.85	269	trace →
11	Claude Opus 4.7 (xhigh)	97.9%	$1.88	70	trace →
12	Gemini 3.6 Flash	97.7%	$2.43	89	trace →
13	GPT-5.6 Sol (xhigh)	96.9%	$4.44	32	trace →
14	GPT-5.6 Sol	96.6%	$0.63	11	trace →
15	Claude Opus 4.7	96.6%	$1.32	56	trace →
16	Claude Sonnet 4.6	96.6%	$10.47	347	trace →
17	GPT 5.5	96.4%	$1.22	17	trace →
18	Muse Spark 1.2 (xhigh)	96.2%	$0.14	88	trace →
19	Claude Fable 5.1 (xhigh)	95.0%	$42.67	337	trace →
20	Gemini 3.7 Flash	91.4%	$2.41	96	trace →
21	Claude Haiku 4.5	88.7%	$0.52	123	trace →
22	Gemini 3.5 Flash	79.0%	$3.53	99	trace →
23	GPT 5.4	76.7%	$0.17	8	trace →
24	GPT 5.4 mini	69.8%	$0.03	13	trace →
25	GPT 6 Astra (max)	66.5%	$11.72	54	trace →
26	GPT 5.5 (xhigh)	64.8%	$7.36	53	trace →
27	Gemini 3.1 Pro	59.7%	$0.88	70	trace →
28	GPT 5 mini	51.4%	$0.02	10	trace →
29	Gemini 3 Flash	13.8%	$0.24	87	trace →

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