BLAKE3-team/BLAKE3 | ProgramBench

the official Rust and C implementations of the BLAKE3 cryptographic hash function

647

Generated Behavioral Tests

100.0%

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)	100.0%	$10.08	102	trace →
2	GPT 6 Astra (max)	100.0%	$5.39	36	trace →
3	Muse Wonder Full	99.7%	TBA	145	trace →
4	Muse Wonder (xhigh)	99.7%	$14.45	225	trace →
5	Claude Fable 5.1 (xhigh)	99.5%	$22.65	170	trace →
6	Claude Opus 5 (xhigh)	98.9%	$36.57	205	trace →
7	Muse Spark 1.3 (xhigh)	98.9%	$0.76	146	trace →
8	GPT-5.6 Sol (xhigh)	98.1%	$4.58	35	trace →
9	Claude Opus 4.8 (xhigh)	98.1%	$8.26	93	trace →
10	Gemini 3.5 Flash	98.1%	$3.59	145	trace →
11	GPT 5.5 (xhigh)	97.8%	$5.97	49	trace →
12	Claude Opus 4.7	97.5%	$3.88	84	trace →
13	Claude Opus 4.7 (xhigh)	96.8%	$6.88	119	trace →
14	Muse Spark 1.1 (xhigh)	96.3%	$0.26	145	trace →
15	Muse Spark 1.2 (xhigh)	96.1%	$0.14	87	trace →
16	GLM-5.2	95.8%	$7.54	183	trace →
17	GPT 5.5 (high)	95.1%	$2.65	34	trace →
18	Claude Opus 4.6	94.9%	$5.68	171	trace →
19	Gemini 3.7 Flash	94.7%	$1.00	78	trace →
20	Gemini 3.6 Flash	93.7%	$2.17	111	trace →
21	GPT 5.5	92.4%	$1.06	15	trace →
22	Claude Sonnet 4.6	92.3%	$4.51	213	trace →
23	Gemini 3.1 Pro	90.6%	$2.22	103	trace →
24	GPT-5.6 Sol	88.4%	$0.44	9	trace →
25	Gemini 3 Flash	51.3%	$0.47	88	trace →
26	Muse Spark 1.3 (max)	32.8%	$8.90	365	trace →
27	GPT 5.4	11.9%	$0.23	10	trace →
28	GPT 5 mini	9.1%	$0.03	18	trace →
29	Claude Haiku 4.5	8.3%	$0.99	148	trace →
30	GPT 5.4 mini	0.2%	$0.05	16	trace →

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