lz4/lz4 | ProgramBench

Extremely Fast Compression algorithm

1,496

Generated Behavioral Tests

99.3%

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)	99.3%	$25.54	175	trace →
2	Muse Wonder Full	97.4%	TBA	375	trace →
3	GPT 6 Astra (max)	96.5%	$12.41	81	trace →
4	Muse Wonder (xhigh)	96.0%	$38.23	422	trace →
5	Muse Spark 1.3 (max)	95.9%	$6.62	365	trace →
6	Claude Fable 5.1 (xhigh)	92.7%	$55.99	253	trace →
7	Claude Opus 5 (xhigh)	92.6%	$30.91	198	trace →
8	Muse Spark 1.3 (xhigh)	91.1%	$1.94	250	trace →
9	GPT 5.5 (xhigh)	88.0%	$8.10	84	trace →
10	Gemini 3.7 Flash	84.9%	$1.32	86	trace →
11	Claude Opus 4.8 (xhigh)	84.5%	$18.31	120	trace →
12	Muse Spark 1.2 (xhigh)	83.8%	$2.76	280	trace →
13	GPT 5.5 (high)	82.7%	$2.92	42	trace →
14	Muse Spark 1.1 (xhigh)	82.7%	$0.97	159	trace →
15	Claude Sonnet 4.6	82.7%	$25.44	501	trace →
16	Gemini 3.6 Flash	81.3%	$2.41	93	trace →
17	Claude Opus 4.7 (xhigh)	80.6%	$4.56	100	trace →
18	Claude Opus 4.7	80.0%	$5.56	155	trace →
19	GPT-5.6 Sol (xhigh)	79.9%	$3.51	47	trace →
20	GPT-5.6 Sol	79.0%	$0.70	12	trace →
21	Claude Opus 4.6	78.7%	$9.86	216	trace →
22	GLM-5.2	76.0%	$15.54	146	trace →
23	Gemini 3.5 Flash	74.9%	$3.79	169	trace →
24	GPT 5.5	74.5%	$0.88	15	trace →
25	GPT 5.4	46.6%	$0.24	10	trace →
26	Gemini 3.1 Pro	36.4%	$0.84	57	trace →
27	Claude Haiku 4.5	22.1%	$0.84	97	trace →
28	GPT 5 mini	14.4%	$0.01	9	trace →
29	GPT 5.4 mini	9.6%	$0.04	10	trace →
30	Gemini 3 Flash	5.2%	$0.20	46	trace →

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