Our users typically think of quality along two axes: content coverage and accuracy of metadata.
You can read more here about what content is covered in OpenAlex here, but the tl;dr is that we generally cover everything other databases cover plus quite a bit more across different output types, languages, and areas of scholarship.
In terms of accuracy of metadata, bibliometric experts around the world reliably use OpenAlex for their research. Studies investigating metadata quality in OpenAlex have found that for works that appear in all databases, metadata quality is comparable (link to Culbert et al. 2024) and also that replicated analyses across databases provide largely identical results (e.g., Alperin et al. 2024). And that's why our data is now used in three major University ranking exercises: CWTS Open Leiden Rankings; Financial Times Business School Rankings; and THE Interdisciplinary Science Rankings.
As we continue to develop OpenAlex and add more sources of data, metadata accuracy will remain a foundational priority. No database is perfect though and all rely on curation from their communities of users. Like the other databases, we also allow users to fix metadata errors they observe in OpenAlex (more on that here). The difference with OpenAlex is that once an error is fixed, it remains open and can be propagated through all other systems. This is in stark contrast to community curation in proprietary databases, which stays within the host system requiring users to curate multiple systems.