Trouble Using o1-pro and o3-pro Models in the Playground

·Jul 25, 2025 07:15 AM

Hey Team, I just got started with the platform and was exploring the playground but was not able to use o1-pro or o3-pro model. I even configured my keys properly it shows : Error code: 404 - {'error': {'message': 'This model is only supported in v1/responses and not in v1/chat/completions.', 'type': 'invalid_request_error', 'param': 'model', 'code': None}}

5 comments

· Sorted by Oldest

RunLLM
APP
·
Support for OpenAI's o1-pro and o3-pro models in the Phoenix playground was added in recent updates, but earlier versions did not support these models. If you are unable to use o1-pro or o3-pro, ensure you are running the latest version of Phoenix, as this feature was added via a recent update (see release details). If you are still experiencing issues after updating, please provide the error message for further troubleshooting.Would you like more detail on how to update or configure your Phoenix instance for these models?Sources: - feat: Update openai reasoning models
💡 Hint: Mention RunLLM in the thread for followups.
Yash M.
·
RunLLM there is error in the configuration please read my entire message
RunLLM
APP
·
Your message appears to be incomplete, and I do not see the full error details or configuration output. To assist you, please provide the exact error message or a screenshot of what you see in the playground when trying to use o1-pro or o3-pro. This will help identify whether the issue is with API keys, model selection, or another configuration step.Could you please paste the full error message or describe what happens when you attempt to use these models?Sources: - https://github.com/Arize-ai/phoenix/issues/8631 - https://github.com/Arize-ai/phoenix/issues/5162
💡 Hint: Mention RunLLM in the thread for followups.
Yash M.
·
Error code: 404 - {'error': {'message': 'This model is only supported in v1/responses and not in v1/chat/completions.', 'type': 'invalid_request_error', 'param': 'model', 'code': None}} RunLLM
Mikyo
·
Hey Yash M. - as the error is indicating, we use chat completions on the playground because it's mainly meant to be a stateless invocation of an LLM but it might be the case (new to me) that 01-pro and 03-pro are stateful via the responses API. I think we definitely need to support responses going forward if this is a case but we would have to adjust a bit. Please feel free to raise a github ticket. We'll post one here if we file one ourselves. Thanks for trying out the playground and feedback!

RunLLM
APP
·
Support for OpenAI's o1-pro and o3-pro models in the Phoenix playground was added in recent updates, but earlier versions did not support these models. If you are unable to use o1-pro or o3-pro, ensure you are running the latest version of Phoenix, as this feature was added via a recent update (see release details). If you are still experiencing issues after updating, please provide the error message for further troubleshooting.Would you like more detail on how to update or configure your Phoenix instance for these models?Sources: - feat: Update openai reasoning models
💡 Hint: Mention RunLLM in the thread for followups.
Yash M.
·
RunLLM there is error in the configuration please read my entire message
RunLLM
APP
·
Your message appears to be incomplete, and I do not see the full error details or configuration output. To assist you, please provide the exact error message or a screenshot of what you see in the playground when trying to use o1-pro or o3-pro. This will help identify whether the issue is with API keys, model selection, or another configuration step.Could you please paste the full error message or describe what happens when you attempt to use these models?Sources: - https://github.com/Arize-ai/phoenix/issues/8631 - https://github.com/Arize-ai/phoenix/issues/5162
💡 Hint: Mention RunLLM in the thread for followups.
Yash M.
·
Error code: 404 - {'error': {'message': 'This model is only supported in v1/responses and not in v1/chat/completions.', 'type': 'invalid_request_error', 'param': 'model', 'code': None}} RunLLM
Mikyo
·
Hey Yash M. - as the error is indicating, we use chat completions on the playground because it's mainly meant to be a stateless invocation of an LLM but it might be the case (new to me) that 01-pro and 03-pro are stateful via the responses API. I think we definitely need to support responses going forward if this is a case but we would have to adjust a bit. Please feel free to raise a github ticket. We'll post one here if we file one ourselves. Thanks for trying out the playground and feedback!