trasta298/keifu | ProgramBench

Git genealogy, untangled. A TUI for navigating commit graphs with color and clarity.

262

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%	$40.26	411	trace →
2	Claude Fable 5.1 (xhigh)	99.6%	$87.16	379	trace →
3	Muse Wonder Full	99.2%	TBA	388	trace →
4	Muse Wonder (xhigh)	98.5%	$37.18	380	trace →
5	GPT 6 Astra (max)	97.3%	$8.84	53	trace →
6	Claude Opus 4.8 (xhigh)	94.7%	$30.71	173	trace →
7	Gemini 3.7 Flash	94.7%	$8.71	169	trace →
8	Muse Spark 1.3 (max)	91.2%	$5.75	375	trace →
9	Muse Spark 1.3 (xhigh)	88.5%	$1.02	192	trace →
10	GPT 5.5 (xhigh)	87.0%	$7.34	70	trace →
11	GPT-5.6 Sol (xhigh)	85.1%	$6.16	43	trace →
12	Gemini 3.6 Flash	82.8%	$4.24	106	trace →
13	Muse Spark 1.1 (xhigh)	82.1%	$0.83	135	trace →
14	Muse Spark 1.2 (xhigh)	77.9%	$0.08	85	trace →
15	GPT-5.6 Sol	77.5%	$1.53	22	trace →
16	GPT 5.5 (high)	75.2%	$3.94	50	trace →
17	Gemini 3.5 Flash	74.4%	$8.57	123	trace →
18	GLM-5.2	68.3%	$35.94	279	trace →
19	Claude Opus 4.7	67.2%	$1.11	48	trace →
20	GPT 5.5	66.4%	$1.02	18	trace →
21	Claude Opus 4.7 (xhigh)	65.6%	$3.33	79	trace →
22	GPT 5.4	36.6%	$0.14	9	trace →
23	Gemini 3 Flash	26.0%	$0.28	64	trace →
24	GPT 5.4 mini	21.8%	$0.10	142	trace →
25	GPT 5 mini	16.0%	$0.01	9	trace →
26	Claude Haiku 4.5	9.9%	$0.64	109	trace →
27	Gemini 3.1 Pro	7.3%	$2.32	92	trace →
28	Claude Opus 5 (xhigh)	0.0%	$57.50	296	trace →
29	Claude Opus 4.6	0.0%	$2.19	178	trace →
30	Claude Sonnet 4.6	0.0%	$5.25	270	trace →

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