XAMPPRocky/tokei | ProgramBench

Count your code, quickly.

732

Generated Behavioral Tests

95.5%

Best Score

Claude Opus 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)

25 runs

#	Model	Score	Cost	Calls
1	Claude Opus 5 (xhigh)	95.5%	$203.20	617	trace →
2	Muse Spark 1.3 (max)	87.2%	$18.46	492	trace →
3	Claude Opus 4.8 (xhigh)	86.9%	$25.80	160	trace →
4	Gemini 3.7 Flash	81.4%	$4.05	209	trace →
5	Gemini 3.5 Flash	75.5%	$7.17	215	trace →
6	Gemini 3.6 Flash	70.4%	$4.33	176	trace →
7	Claude Sonnet 4.6	69.5%	$21.30	409	trace →
8	Claude Opus 4.7 (xhigh)	62.6%	$11.82	171	trace →
9	Muse Spark 1.1 (xhigh)	61.1%	$0.49	219	trace →
10	GPT 5.5 (high)	60.9%	$2.96	27	trace →
11	GPT-5.6 Sol	60.9%	$1.07	17	trace →
12	GLM-5.2	59.7%	$43.16	252	trace →
13	Muse Spark 1.2 (xhigh)	57.7%	$0.97	135	trace →
14	Claude Opus 4.7	54.8%	$3.25	95	trace →
15	Gemini 3.1 Pro	53.4%	$1.81	112	trace →
16	GPT 5.5	51.0%	$1.57	19	trace →
17	Gemini 3 Flash	40.4%	$0.26	71	trace →
18	GPT 5.4	32.0%	$0.39	11	trace →
19	Claude Haiku 4.5	27.0%	$0.71	94	trace →
20	GPT-5.6 Sol (xhigh)	25.5%	$4.88	41	trace →
21	Muse Spark 1.3 (xhigh)	18.6%	$0.95	173	trace →
22	GPT 5 mini	12.2%	$0.01	8	trace →
23	GPT 5.4 mini	8.1%	$0.04	8	trace →
24	GPT 5.5 (xhigh)	2.0%	$10.01	98	trace →
25	Claude Opus 4.6	0.0%	$2.82	193	trace →

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