Esubaalew/run | ProgramBench

Universal multi-language runner and smart REPL written in Rust.

1,212

Generated Behavioral Tests

99.6%

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)

29 runs

#	Model	Score	Cost	Calls
1	Claude Opus 5.5 (xhigh)	99.6%	$27.33	370	trace →
2	Muse Wonder (xhigh)	98.1%	$32.37	346	trace →
3	GPT 6 Astra (max)	98.1%	$12.09	57	trace →
4	Muse Wonder Full	97.6%	TBA	383	trace →
5	Claude Opus 5 (xhigh)	92.8%	$45.41	334	trace →
6	GLM-5.2	88.2%	$36.85	234	trace →
7	Claude Opus 4.7 (xhigh)	88.0%	$19.96	282	trace →
8	Claude Opus 4.8 (xhigh)	86.5%	$14.03	122	trace →
9	Claude Opus 4.6	85.2%	$14.72	342	trace →
10	Claude Sonnet 4.6	85.2%	$32.29	669	trace →
11	GPT 5.5 (xhigh)	84.5%	$6.28	78	trace →
12	Muse Spark 1.2 (xhigh)	82.4%	$0.17	143	trace →
13	GPT-5.6 Sol	81.3%	$0.76	11	trace →
14	GPT 5.5 (high)	80.6%	$3.33	44	trace →
15	Muse Spark 1.1 (xhigh)	80.0%	$0.24	98	trace →
16	Claude Opus 4.7	73.2%	$1.91	73	trace →
17	GPT 5.5	72.7%	$0.82	14	trace →
18	Muse Spark 1.3 (max)	70.4%	$9.18	323	trace →
19	GPT 5.4	70.1%	$0.21	11	trace →
20	Muse Spark 1.3 (xhigh)	69.0%	$1.95	255	trace →
21	GPT-5.6 Sol (xhigh)	67.7%	$5.47	39	trace →
22	Gemini 3.7 Flash	64.7%	$1.88	119	trace →
23	Gemini 3.6 Flash	63.5%	$2.98	105	trace →
24	Gemini 3.5 Flash	59.7%	$4.26	156	trace →
25	Gemini 3.1 Pro	58.5%	$3.42	156	trace →
26	Gemini 3 Flash	54.5%	$0.26	85	trace →
27	Claude Haiku 4.5	42.3%	$0.32	104	trace →
28	GPT 5.4 mini	37.6%	$0.03	18	trace →
29	GPT 5 mini	1.6%	$0.03	16	trace →

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