Hey everyone! I'm experiencing some sluggishness with the o4-mini model while using the Azure AI Foundry playground. Even for simple prompts like 'What is the difference between power and authority?', I find that it takes around 2 minutes to get a response, which is just a chain of thoughts rather than a complete answer. Is there a way to speed up the response time? Any tips would be appreciated, thanks!
3 Answers
Have you checked what deployment model you're using for o4-mini? Sometimes, the settings can really affect response times.
I've noticed similar issues lately, especially when hosted in Sweden Central. It seems like o3 is also running slow and often gives errors. I had some success by lowering the max completion tokens, but it's still not very reliable.
If you change your deployment region from Sweden Central to EastUS2, you might find that the response times improve significantly. I tried it and it worked better for me!
Yeah, we're definitely using o4-mini, version 2025-04-16.