openai: Unsupported parameter: 'max_tokens' is not supported with this model. Use 'max_completion_tokens' instead.

[Fristi]

When using a GPT 5+ model I encounter this

What should we do to remediate this ?

[adamw]

This isn't a defect in sttp-ai. OpenAI's GPT-5 (and other reasoning models) reject the max_tokens parameter and require max_completion_tokens instead — this is an OpenAI API constraint, not a library bug. sttp-ai's ChatBody (openai/src/main/scala/sttp/ai/openai/requests/completions/chat/ChatRequestBody.scala) already exposes both fields as independent optional parameters: maxTokens (serialized as max_tokens) and maxCompletionTokens (serialized as max_completion_tokens). They are serialized separately and null values are dropped, so only the field you set is sent. Remediation: when targeting a GPT-5+ model, leave maxTokens = None and set maxCompletionTokens instead, e.g. ChatBody(messages = ..., model = ChatCompletionModel.GPT5, maxCompletionTokens = Some(1000)).

[Fristi]

Thanks! Could add this to the docs ? :)