BurntSushi/xsv | ProgramBench

A fast CSV command line toolkit written in Rust.

1,182

Generated Behavioral Tests

99.8%

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)	99.8%	$18.33	197	trace →
2	Claude Fable 5.1 (xhigh)	99.7%	$83.43	272	trace →
3	GPT 6 Astra (max)	99.6%	$8.46	51	trace →
4	Muse Wonder Full	98.9%	TBA	297	trace →
5	Muse Wonder (xhigh)	98.7%	$19.05	265	trace →
6	Claude Opus 5 (xhigh)	98.5%	$101.89	334	trace →
7	Claude Opus 4.8 (xhigh)	94.1%	$23.30	156	trace →
8	GPT 5.5 (xhigh)	90.7%	$11.17	93	trace →
9	Gemini 3.5 Flash	88.4%	$7.21	198	trace →
10	GPT-5.6 Sol (xhigh)	88.0%	$6.72	43	trace →
11	Gemini 3.7 Flash	87.5%	$1.85	137	trace →
12	GPT 5.5 (high)	87.1%	$2.70	40	trace →
13	Claude Opus 4.7 (xhigh)	85.8%	$24.54	350	trace →
14	Muse Spark 1.2 (xhigh)	85.5%	$1.37	183	trace →
15	Claude Opus 4.6	82.7%	$17.25	288	trace →
16	Claude Opus 4.7	82.0%	$4.51	106	trace →
17	Gemini 3.6 Flash	81.4%	$3.39	131	trace →
18	Muse Spark 1.3 (xhigh)	81.3%	$1.09	229	trace →
19	Claude Sonnet 4.6	80.6%	$15.53	367	trace →
20	Muse Spark 1.1 (xhigh)	77.4%	$0.40	193	trace →
21	GPT 5.5	77.2%	$1.18	21	trace →
22	GPT-5.6 Sol	70.7%	$0.64	10	trace →
23	GPT 5.4	65.6%	$0.32	8	trace →
24	GLM-5.2	49.9%	$32.01	327	trace →
25	Gemini 3 Flash	39.6%	$0.33	97	trace →
26	Gemini 3.1 Pro	32.2%	$1.10	78	trace →
27	GPT 5.4 mini	6.3%	$0.02	11	trace →
28	Muse Spark 1.3 (max)	2.7%	$1.98	172	trace →
29	GPT 5 mini	2.7%	$0.03	18	trace →
30	Claude Haiku 4.5	0.0%	$0.89	137	trace →

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