chore: Update default OpenAI model to gpt-5-mini

[vblagoje]

Why

Updates the default OpenAI model from gpt-4o-mini to gpt-5-mini to use the latest model version by default.

What

• Changed default model parameter in OpenAIChatGenerator and OpenAIGenerator from gpt-4o-mini to gpt-5-mini
• Updated docstrings and documentation examples to reflect the new default
• Updated test expectations to match the new default model
• Made tests more robust by avoiding hardcoded model names where appropriate

How can it be used

The default model is now gpt-5-mini when initializing OpenAI generators without specifying a model:

# Uses gpt-5-mini by default
generator = OpenAIChatGenerator()

# Still works with explicit model specification
generator = OpenAIChatGenerator(model="gpt-4o")

How did you test it

• Updated all unit tests to expect gpt-5-mini as the default
• Made integration tests more flexible to handle model name variations
• Updated test mocks and fixtures to use the new default
• Marked two flaky integration tests with retry decorators

Notes for the reviewer

This is a straightforward model update. All tests pass and the change is backward compatible - users can still specify any model explicitly.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment

Project	Deployment	Preview	Comments	Updated (UTC)
haystack-docs	Ignored	Preview		Dec 1, 2025 0:37am

[coveralls]

Pull Request Test Coverage Report for Build 19822905143

Details

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage remained the same at 92.189%

---

Totals
Change from base Build 19820645949:	0.0%
Covered Lines:	14033
Relevant Lines:	15222

---

💛 - Coveralls

[mpangrazzi]

Looks ok!

[sjrl]

@vblagoje it still may be good to label this as a breaking change and in the release notes add an upgrade section. Btw when is the end of life of gpt-4o?

[vblagoje]

@vblagoje it still may be good to label this as a breaking change and in the release notes add an upgrade section. Btw when is the end of life of gpt-4o?

ok will do, chat version of it very soon but not the gpt-4o - seems like it will be around still for some time

[mpangrazzi]

@vblagoje FYI there was a (probably) fake tests. Re-running test "fixed" it.

[vblagoje]

@sjrl @mpangrazzi I'll integrate this one now and then issue PR for Azure and finally huggingface, lmk if you have some objections

[sjrl]

@sjrl @mpangrazzi I'll integrate this one now and then issue PR for Azure and finally huggingface, lmk if you have some objections

be careful for the Azure one, it's possible that we may want to leave gpt-4o as default there since Azure prefers users to use the responses api to use gpt-5

[vblagoje]

@sjrl @mpangrazzi I'll integrate this one now and then issue PR for Azure and finally huggingface, lmk if you have some objections

be careful for the Azure one, it's possible that we may want to leave gpt-4o as default there since Azure prefers users to use the responses api to use gpt-5

Aha, ok, let's go with 4.1-mini there then?