php/php-src | ProgramBench

The PHP Interpreter

14,288

Generated Behavioral Tests

33.3%

Best Score

Claude Fable 5.1 (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)

29 runs

#	Model	Score	Cost	Calls
1	Claude Fable 5.1 (xhigh)	33.3%	$98.42	316	trace →
2	Muse Wonder (xhigh)	26.5%	$35.90	348	trace →
3	Claude Opus 4.8 (xhigh)	17.8%	$27.39	167	trace →
4	GPT-5.6 Sol (xhigh)	16.1%	$6.78	34	trace →
5	Muse Spark 1.3 (max)	15.0%	$7.32	338	trace →
6	GPT 5.5 (xhigh)	13.2%	$12.89	115	trace →
7	Muse Wonder Full	12.4%	TBA	544	trace →
8	Gemini 3.7 Flash	12.3%	$1.00	57	trace →
9	GPT 6 Astra (max)	12.3%	$16.44	56	trace →
10	GPT-5.6 Sol	8.6%	$1.05	13	trace →
11	GPT 5.5 (high)	8.0%	$3.23	36	trace →
12	Muse Spark 1.2 (xhigh)	5.3%	$2.91	252	trace →
13	Claude Haiku 4.5	4.8%	$0.52	119	trace →
14	Claude Opus 4.7 (xhigh)	3.2%	$3.26	73	trace →
15	GPT 5.5	3.1%	$1.47	16	trace →
16	GPT 5.4	2.2%	$0.24	7	trace →
17	Gemini 3.1 Pro	1.7%	$1.33	109	trace →
18	GPT 5.4 mini	1.1%	$0.03	13	trace →
19	Muse Spark 1.3 (xhigh)	1.0%	$1.11	187	trace →
20	GPT 5 mini	0.9%	$0.04	14	trace →
21	Claude Opus 4.7	0.9%	$0.25	14	trace →
22	Gemini 3 Flash	0.6%	$0.22	65	trace →
23	Gemini 3.6 Flash	0.3%	$3.46	132	trace →
24	Claude Opus 5.5 (xhigh)	0.0%	$62.92	245	trace →
25	Claude Opus 5 (xhigh)	0.0%	$1.41	24	trace →
26	GLM-5.2	0.0%	$0.39	31	trace →
27	Gemini 3.5 Flash	0.0%	$3.13	124	trace →
28	Claude Opus 4.6	0.0%	$2.94	205	trace →
29	Claude Sonnet 4.6	0.0%	$7.38	326	trace →

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