The State of OpenAPI 3.2 Support

About this paper

OpenAPI 3.2 shipped on 19 September 2025 with features providers had wanted for years: a QUERY method, schemas for streamed responses, hierarchical tags, and a way for a document to name itself. A specification is only as useful as the tools that read it, so API Evangelist research tested 133 of them against the same contracts. Fifty tools were run directly, and eighty-seven hosted products and frameworks were read from their own documentation, changelogs and source.

One year in, a quarter of the market handles a 3.2 document, and only fifteen tools say so themselves. Partial support is the most common real state. A small number of shared parsers decide what whole branches of the market can read, and most hosted vendors do not mention 3.2 at all. This paper names who supports what, explains why OpenAPI 3.1 is still unfinished five years on, and gives API providers and tooling vendors a plain plan for what to do next.

What's inside

1. What 3.2 adds, and why it is worth wanting
2. How the census was measured
3. The numbers, by class of tool
4. The fifteen that wrote it down
5. Partial is the market
6. Where 3.2 stops by name
7. The silence
8. 3.1 is not finished, and frameworks decide
9. The neighbours: AsyncAPI, Arazzo and Overlay
10. What to do
11. Anti-patterns worth watching for
12. Self-assessment
13. Where this is going

What you get for $25.00