Results

condor and vulture 2.16 run on the same real projects.

Findings

Fewer findings means fewer false positives: framework code, model fields and stub parameters are no longer reported.

Project	What it is	vulture	condor	Change
vulture	library	0	0	—
full-stack-fastapi-template	FastAPI application	50	26	−48%
httpx	library, uses all	63	7	−89%
requests	library	93	93	same
click	library	119	58	−51%
attrs	library, uses all	119	78	−34%
rich	library	130	113	−13%
flask	library / web framework	333	81	−76%
scrapy	application framework	438	422	−4%
celery	application framework	875	798	−9%
pydantic	library	1,362	514	−62%
fastapi	web framework + many FastAPI apps in docs_src	1,664	177	−89%
black	application	1,793	1,775	−1%
sqlalchemy	large library	2,528	1,968	−22%
netbox	Django application	2,912	2,116	−27%
django	large web framework	3,180	2,146	−33%
saleor	large Django application	3,263	1,478	−55%
typeshed	stub files	3,709	1,566	−58%
cpython-stdlib	CPython 3.14 standard library and its tests	5,610	5,266	−6%
Total		28,241	18,682	−34%

Findings reported with default options. Exit codes: identical in all 79 runs, including the other option sets.

Speed

Machine: Apple M2, 8 cores, Darwin. vulture 2.16 on CPython 3.14.4; condor release build. Mean wall-clock time over several runs after two warm-up runs (hyperfine), so the file system cache is warm. Same corpus revisions as REPORT.md. Load average when the run ended: about 23 on 8 cores (busy machine: timings are slower than usual for both tools).

Input	.py files	vulture (s)	condor (s)	Speed-up
one small file	1	0.051 ± 0.001	0.004 ± 0.000	13×
requests	36	0.173 ± 0.001	0.022 ± 0.001	8×
flask	82	0.248 ± 0.003	0.030 ± 0.001	8×
rich	190	0.456 ± 0.014	0.045 ± 0.001	10×
fastapi	1283	1.065 ± 0.011	0.124 ± 0.001	9×
django	2788	6.733 ± 1.575	0.368 ± 0.003	18×
saleor	3524	11.604 ± 3.022	0.624 ± 0.056	19×

Every run, option set and sample is in the full report.