rust-lang/mdBook | ProgramBench

Create book from markdown files. Like Gitbook but implemented in Rust

1,114

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)

30 runs

#	Model	Score	Cost	Calls
1	Claude Opus 5.5 (xhigh)	99.6%	$58.02	397	trace →
2	Claude Fable 5.1 (xhigh)	98.9%	$102.67	457	trace →
3	Muse Wonder Full	95.2%	TBA	465	trace →
4	Muse Wonder (xhigh)	94.1%	$40.88	388	trace →
5	Claude Opus 5 (xhigh)	94.0%	$119.20	481	trace →
6	GPT 6 Astra (max)	93.2%	$11.60	46	trace →
7	GPT-5.6 Sol (xhigh)	76.5%	$7.10	48	trace →
8	Muse Spark 1.3 (xhigh)	72.6%	$2.23	295	trace →
9	GPT 5.5 (xhigh)	72.5%	$11.95	87	trace →
10	Muse Spark 1.2 (xhigh)	63.7%	$0.94	114	trace →
11	Claude Opus 4.8 (xhigh)	63.6%	$24.82	184	trace →
12	GPT 5.5 (high)	61.8%	$3.88	33	trace →
13	Muse Spark 1.1 (xhigh)	60.6%	$0.45	191	trace →
14	GPT-5.6 Sol	60.1%	$1.15	18	trace →
15	GLM-5.2	59.4%	$62.65	319	trace →
16	Claude Sonnet 4.6	55.5%	$50.89	644	trace →
17	Gemini 3.6 Flash	53.8%	$3.27	140	trace →
18	GPT 5.5	48.8%	$1.10	15	trace →
19	GPT 5.4	40.8%	$0.26	9	trace →
20	Claude Opus 4.7	37.9%	$1.52	60	trace →
21	Gemini 3.5 Flash	34.8%	$2.12	74	trace →
22	Claude Opus 4.7 (xhigh)	31.2%	$5.72	155	trace →
23	Gemini 3 Flash	25.6%	$0.25	80	trace →
24	Gemini 3.1 Pro	21.5%	$1.83	110	trace →
25	Claude Haiku 4.5	18.2%	$0.74	139	trace →
26	GPT 5 mini	8.6%	$0.04	18	trace →
27	Gemini 3.7 Flash	4.1%	$1.24	116	trace →
28	Claude Opus 4.6	4.1%	$13.14	244	trace →
29	Muse Spark 1.3 (max)	0.0%	$9.38	371	trace →
30	GPT 5.4 mini	0.0%	$0.03	9	trace →

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