max_tokens is not supported with this model max_completion_tokens

Not specified by the original author.

TL;DR: Reasoning models reject maxtokens outright, and the replacement has a trap: it budgets hidden reasoning tokens too. Send maxcompletiontokens everywhere, size it generously, and recognize the empty-answer signature of a starved budget. Use 'maxcompletiontokens' instead. Send maxcompletiontokens everywhere, never maxtokens.

The symptom

400 Unsupported parameter: 'max_tokens' is not supported with this model. Use 'max_completion_tokens' instead. Fires on o-series and GPT-5 reasoning models. Code that worked on GPT-4o breaks the moment it points at a reasoning model.

Confirm the cause

max_completion_tokens replaced max_tokens for newer models, and the rename is semantic: reasoning models generate internal reasoning tokens that are billed but never returned, so "tokens generated" and "tokens you see back" are no longer the same number. The new parameter works on all current models, making it the portable choice.

The trap: the budget covers ALL output tokens including hidden reasoning. A small budget can be eaten entirely by internal thinking, leaving visible content empty with finish_reason: "length". An empty answer with finish_reason length on a reasoning model usually means the budget starved, not that the model had nothing to say.

The fix

1. Send max_completion_tokens everywhere, never max_tokens. One code path covers GPT-4o, GPT-3.5-era, and reasoning models.
2. Size it generously for reasoning models. Empty content plus finishreason length means raise the budget substantially or lower `reasoningeffort` to cap thinking overhead.
3. Grep every call site when migrating, including wrappers and SDK defaults. The ones you miss 400 at 2am.

Verify the fix

Call the reasoning model with the new parameter and confirm 200. Deliberately test the edge: a tiny budget should produce the empty-content plus finish_reason length signature, and your code should read that as "budget too small", not "model broken".

When to use this

• This covers exactly what the title says: OpenAI 400 Unsupported parameter max_tokens.
• You are setting this up for the first time, or auditing an existing setup.
• You want the key gotchas in one place before you start.

When not to use this

• You are doing a different workflow with OpenAI; these steps are specific to the title above.
• You need the full reference docs; this is the short path, not the manual.

Compatibility

• Mentions 3.5 in the original; behavior may differ on much older or newer releases.

Reasoning models return 400 Unsupported parameter for maxtokens with the instruction to use maxcompletiontokens, which works across current models and budgets all output tokens including hidden reasoning tokens. The reasoning guide shows reasoningtokens reported under usage outputtokensdetails: billed output that never appears in the response (https://platform.openai.com/docs/guides/error-codes, https://platform.openai.com/docs/guides/reasoning).

Not specified by the original author.

TL;DR

Use 'maxcompletiontokens' instead." Semantic Kernel automatically passed max\_tokens, which o3 models reject.

Steps

1. Invoking Semantic Kernel Python's AzureChatCompletion against an o3-series reasoning model failed with "Unsupported parameter: 'maxtokens' is not supported with this model. Use 'maxcompletiontokens' instead." Semantic Kernel automatically passed `maxtokens`, which o3 models reject. The user expected the standard prompt execution settings to just work with reasoning models.
2. Configure max_completion_tokens instead of max_tokens: in SK Python, OpenAIChatPromptExecutionSettings (the base class for OpenAI prompt execution settings) has a max_completion_tokens attribute, and the child AzureChatPromptExecutionSettings adds extra_body if needed. Setting that value works fine with reasoning models, per maintainer moonbox3 in the thread.

When to use

You are seeing this: Invoking Semantic Kernel Python's AzureChatCompletion against an o3-series reasoning model failed with "Unsupported parameter: 'maxtokens' is not supported with this model. Use this skill when you run into "Semantic Kernel Python + Azure OpenAI o3: Unsupported parameter maxtokens, use maxcompletiontokens".

When not to use

If your error message or symptom does not match what is described above, this is probably not your fix. Search for your exact error text instead of forcing this one to fit.

Versions

No specific versions are mentioned in the source material, so treat the fix as generally applicable and check the examples against whatever you have installed.

Why this happens

The original report does not dig into a root cause. It documents the symptom and the fix that resolved it.

Question asked in the thread: Invoking Semantic Kernel Python's AzureChatCompletion against an o3-series reasoning model failed with "Unsupported parameter: 'maxtokens' is not supported with this model. Use 'maxcompletiontokens' instead." Semantic Kernel automatically passed `maxtokens`, which o3 models reject. The user expected the standard prompt execution settings to just work with reasoning models.

Verified answer: Configure max_completion_tokens instead of max_tokens: in SK Python, OpenAIChatPromptExecutionSettings (the base class for OpenAI prompt execution settings) has a max_completion_tokens attribute, and the child AzureChatPromptExecutionSettings adds extra_body if needed. Setting that value works fine with reasoning models, per maintainer moonbox3 in the thread. Source: https://github.com/microsoft/semantic-kernel/issues/12865

Thread: https://vectle.com/threads/thr_ALZMb0ORS9uDsQ0qgvmc3g

Not specified by the original author.

TL;DR

Make sure the text is clear." [5](vscode-notebook-cell:?execution_count=11&line=5) # imgprompt = "Ask me how I'm doing" ----> [7](vscode-notebook-cell:?executioncount=11&line=7) result = dalle.initiatechat(critic, message=imgprompt) AuthenticationError: Error code: 401 - {'error': {'code': 'invalidapikey', 'message': 'Incorrect API key provided: f***************************4e23. @whiskyboy max_token solved the cutoff problem!

[reporter nazkhan-8451]: @whiskyboy max_token solved the cutoff problem! Are you using Azure Dalle? Because for Dalle I get the following error. But the api key is fine because when I forced it to accept it (showed above comment) it works. AuthenticationError Traceback (most recent call last) Cell In[11], [line 7](vscode-notebook-cell:?execution_count=11&line=7) [4](vscode-notebook-cell:?execution_count=11&line=4) img_prompt = "A happy dog wearing a shirt saying 'I Love AutoGen'. Make sure the text is clear." [5](vscode-notebook-cell:?execution_count=11&line=5) # img_prompt = "Ask me how I'm doing" ----> [7](vscode-notebook-cell:?execution_count=11&line=7) result = dalle.initiate_chat(critic, message=img_prompt) AuthenticationError: Error code: 401 - {'error': {'code': 'invalid_api_key', 'message': 'Incorrect API key provided: f***************************4e23. You can find your API key at https://platform.openai.com/account/api-keys.', 'param': None, 'type': 'invalid_request_error'}}

Context: Issue microsoft/autogen#2510 (closed, 29 comments): ### Describe the bug Followed the notebook [https://github.com/microsoft/autogen/blob/main/notebook/agentchat_image_generation_capability.ipynb](url), but getting the following response: [image] Code: ```python import autogen configlistgpt4 = autogen.configlistfromjson( "OAICONFIGLIST", filterdict={ "model": ["wag-gpt4-128k"], }, ) configlistgpt4vision = autogen.configlistfromjson( "OAICONFIGLIST", filter_dict={ "model": ["gpt4-

When to use

• You hit this exact issue with Bug].
• The symptom matches: Dalle-Critic not working.

When NOT to use

• A different error or symptom from Bug]; the cause here is specific to this one.

Compatibility

Reported against Bug].

Edge cases

• If your error message differs even slightly, this is probably a different issue; search the exact text.
• If the fix does not help, capture the full error output and check the source link for updates.

Source: Issue microsoft/autogen#2510 (closed, 29 comments): ### Describe the bug Followed the notebook [https://github.com/microsoft/autogen/blob/main/notebook/agentchat_image_generation_capability.ipynb](url), but getting the following response: [image] Code: ```python import autogen configlistgpt4 = autogen.configlistfromjson( "OAICONFIGLIST", filterdict={ "model": ["wag-gpt4-128k"], }, ) configlistgpt4vision = autogen.configlistfromjson( "OAICONFIGLIST", filter_dict={ "model": ["gpt4-

Verified resolution: [reporter nazkhan-8451]: @whiskyboy max_token solved the cutoff problem! Are you using Azure Dalle? Because for Dalle I get the following error. But the api key is fine because when I forced it to accept it (showed above comment) it works. AuthenticationError Traceback (most recent call last) Cell In[11], [line 7](vscode-notebook-cell:?execution_count=11&line=7) [4](vscode-notebook-cell:?execution_count=11&line=4) img_prompt = "A happy dog wearing a shirt saying 'I Love AutoGen'. Make sure the text is clear." [5](vscode-notebook-cell:?execution_count=11&line=5) # img_prompt = "Ask me how I'm doing" ----> [7](vscode-notebook-cell:?execution_count=11&line=7) result = dalle.initiate_chat(critic, message=img_prompt) AuthenticationError: Error code: 401 - {'error': {'code': 'invalid_api_key', 'message': 'Incorrect API key provided: f***************************4e23. You can find your API key at https://platform.openai.com/account/api-keys.', 'param': None, 'type': 'invalid_request_error'}}

Source URL: https://github.com/microsoft/autogen/issues/2510