tinycc/tinycc | ProgramBench

Unofficial mirror of mob development branch

1,978

Generated Behavioral Tests

80.0%

Best Score

Muse Wonder Full

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	Muse Wonder Full	80.0%	TBA	393	trace →
2	GPT 5.5 (xhigh)	72.4%	$6.30	65	trace →
3	Gemini 3.6 Flash	67.3%	$3.22	150	trace →
4	GPT-5.6 Sol	64.9%	$0.59	10	trace →
5	Gemini 3.5 Flash	59.7%	$2.29	117	trace →
6	Muse Wonder (xhigh)	23.7%	$27.80	327	trace →
7	GPT 6 Astra (max)	21.6%	$19.15	64	trace →
8	Claude Opus 5.5 (xhigh)	21.5%	$58.88	298	trace →
9	Muse Spark 1.2 (xhigh)	20.9%	$0.46	77	trace →
10	Claude Fable 5.1 (xhigh)	18.8%	$93.11	288	trace →
11	GPT 5.5 (high)	17.7%	$3.43	36	trace →
12	GPT 5.5	16.0%	$1.29	17	trace →
13	Muse Spark 1.3 (max)	14.1%	$10.19	421	trace →
14	GPT 5.4	12.8%	$0.43	12	trace →
15	GLM-5.2	11.1%	$19.13	193	trace →
16	GPT-5.6 Sol (xhigh)	10.6%	$6.68	41	trace →
17	Claude Sonnet 4.6	9.2%	$100.12	879	trace →
18	Muse Spark 1.3 (xhigh)	8.8%	$2.05	238	trace →
19	Claude Opus 4.6	8.0%	$12.01	189	trace →
20	Gemini 3 Flash	4.8%	$0.16	50	trace →
21	Claude Opus 4.8 (xhigh)	4.3%	$16.40	119	trace →
22	Gemini 3.1 Pro	3.8%	$1.52	108	trace →
23	Claude Opus 4.7 (xhigh)	3.7%	$0.75	22	trace →
24	Claude Opus 4.7	3.3%	$0.45	22	trace →
25	GPT 5.4 mini	2.2%	$0.02	7	trace →
26	GPT 5 mini	1.6%	$0.02	12	trace →
27	Muse Spark 1.1 (xhigh)	0.8%	$0.70	204	trace →
28	Gemini 3.7 Flash	0.0%	$0.28	63	trace →
29	Claude Haiku 4.5	0.0%	$0.40	108	trace →

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