astaxie/bat | ProgramBench

Go implement CLI, cURL-like tool for humans

1,091

Generated Behavioral Tests

100.0%

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)	100.0%	$13.85	222	trace →
2	Muse Wonder (xhigh)	94.8%	$26.45	307	trace →
3	Muse Wonder Full	91.5%	TBA	289	trace →
4	Claude Opus 5 (xhigh)	89.3%	$35.99	209	trace →
5	Claude Fable 5.1 (xhigh)	89.0%	$25.71	174	trace →
6	GPT 6 Astra (max)	88.5%	$6.42	36	trace →
7	Muse Spark 1.3 (xhigh)	86.3%	$1.28	200	trace →
8	Muse Spark 1.3 (max)	86.1%	$1.17	144	trace →
9	GPT-5.6 Sol (xhigh)	82.7%	$5.35	40	trace →
10	Claude Opus 4.8 (xhigh)	82.4%	$14.77	111	trace →
11	Muse Spark 1.1 (xhigh)	81.0%	$0.24	162	trace →
12	GPT 5.5 (xhigh)	79.5%	$5.83	49	trace →
13	GPT 5.5	79.4%	$1.16	17	trace →
14	Gemini 3.7 Flash	79.1%	$1.43	99	trace →
15	GPT 5.5 (high)	78.2%	$2.19	21	trace →
16	Claude Opus 4.6	71.8%	$9.63	188	trace →
17	GPT-5.6 Sol	71.6%	$0.73	13	trace →
18	Claude Opus 4.7 (xhigh)	70.9%	$13.25	132	trace →
19	Gemini 3.5 Flash	69.9%	$5.97	142	trace →
20	Claude Opus 4.7	66.4%	$4.95	126	trace →
21	Gemini 3 Flash	65.4%	$0.29	85	trace →
22	Claude Sonnet 4.6	64.5%	$18.12	365	trace →
23	Muse Spark 1.2 (xhigh)	63.8%	$0.05	65	trace →
24	Gemini 3.6 Flash	62.1%	$2.88	102	trace →
25	GLM-5.2	61.0%	$12.53	137	trace →
26	GPT 5.4	58.9%	$0.17	9	trace →
27	Gemini 3.1 Pro	50.5%	$1.25	55	trace →
28	Claude Haiku 4.5	49.9%	$0.58	88	trace →
29	GPT 5.4 mini	26.4%	$0.04	8	trace →
30	GPT 5 mini	16.0%	$0.01	9	trace →

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