naggie/dstask | ProgramBench

Git powered terminal-based todo/note manager -- markdown note page per task. Single binary!

1,278

Generated Behavioral Tests

99.3%

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	99.3%	TBA	348	trace →
2	Claude Opus 5.5 (xhigh)	99.0%	$35.19	384	trace →
3	GPT 6 Astra (max)	96.3%	$10.60	58	trace →
4	Muse Wonder (xhigh)	96.1%	$42.27	377	trace →
5	Claude Opus 5 (xhigh)	96.1%	$97.14	455	trace →
6	Claude Fable 5.1 (xhigh)	95.2%	$52.78	363	trace →
7	Muse Spark 1.3 (max)	93.3%	$3.18	216	trace →
8	Gemini 3.7 Flash	82.6%	$4.08	235	trace →
9	Claude Opus 4.8 (xhigh)	82.5%	$20.98	169	trace →
10	GPT-5.6 Sol (xhigh)	82.1%	$5.07	41	trace →
11	GPT 5.5 (high)	76.8%	$4.37	45	trace →
12	Gemini 3.6 Flash	75.9%	$6.51	165	trace →
13	Muse Spark 1.3 (xhigh)	75.3%	$1.95	320	trace →
14	GPT 5.5 (xhigh)	71.3%	$11.41	107	trace →
15	Muse Spark 1.1 (xhigh)	67.4%	$0.68	179	trace →
16	GPT-5.6 Sol	67.3%	$1.31	13	trace →
17	Gemini 3.5 Flash	66.5%	$7.09	198	trace →
18	GLM-5.2	64.5%	$25.06	220	trace →
19	Claude Opus 4.7 (xhigh)	64.1%	$11.13	177	trace →
20	Claude Opus 4.6	58.8%	$10.21	241	trace →
21	Claude Sonnet 4.6	58.2%	$31.15	632	trace →
22	GPT 5.5	55.6%	$1.25	16	trace →
23	Claude Opus 4.7	50.7%	$3.06	96	trace →
24	GPT 5.4	48.8%	$0.49	12	trace →
25	Claude Haiku 4.5	30.8%	$0.69	95	trace →
26	Gemini 3 Flash	22.6%	$0.31	90	trace →
27	GPT 5 mini	7.2%	$0.02	18	trace →
28	Gemini 3.1 Pro	6.1%	$1.52	66	trace →
29	GPT 5.4 mini	1.8%	$0.02	7	trace →
30	Muse Spark 1.2 (xhigh)	0.0%	$0.38	269	trace →

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