Is Realtime API out of BETA now?
Just FYI the python/JS code blocks in the docs here are reversed
Is there a plan to have fine tuning for the audio models? And if so, when can we expect it?
Which tier are you on? o1 has started rolling out to developers on usage tier 5 today, and rollout will complete after a few weeks. Team’s working on adding expanding to more tiers. (You’ll get an email notifying you once it’s available to you in the API and Playground.)
- It’s a cool protocol!
- For DALL-E, thanks for the input!
- I hear you on the vision improvements. We’re working on something in this space now and hope to have some improved vision models soon.
Thank you, valuable info for me and my org!
The keys are locked to a specific Realtime API session, so their use is substantially limited compared to typical OpenAI API keys.
Not sure if this answers your use case, but I’ve built out automations that use multiple separate assistants on a single chat thread. Each automation module is retaining the thread ID, and using a different assistant ID with a prompt to generate a more specific request, with conversation history maintained in the thread.
When will DALL·E’s seed functionality be available via the API?
This feature has been available in ChatGPT for over a year and is essential for developers. DALL·E’s prompt adherence and quality is unmatched, and its seed functionality delivers more consistent results than competitive image generation models. Releasing it via the API would unlock incredible potential for the broader developer community. Please consider prioritizing this! 
I wanted to do a YouTube 2024 rewind video
Preference fine-tuning can encourage longer responses, however we still have a maximum number of output content tokens for each model.
Great question on the developer message! The best resource to read more is our Model Spec detailing the hierarchy:
Follow the chain of command. Subject to its rules, the Model Spec explicitly delegates all remaining power to the developer (for API use cases) and end user. In some cases, the user and developer will provide conflicting instructions; in such cases, the developer message should take precedence.
Thank you for the note on PHP! With the addition of Java and Go today, our goal is to ultimately support the most popular languages and tech stacks. No timeline to share yet for a PHP SDK, but in the meantime, we recommend community-supported libraries to get started.
Thanks that’s helps a lot. Tier 4… This won’t do! Take my money please 
Have a good day and thanks again 
If you wish to discuss a response in detail, please create a new thread with a link to the reply in the body using the 
icon to save filling the AMA thread.
Yes! It is possible to do supervised fine-tuning and then run preference fine-tuning. We have seen good results from running SFT first and then DPO after.
We are actively working on a guide around how to approach reinforcement fine-tuning, and will publish it once it’s ready! But there are still some details we wanted to work out first.
At a high level though, I would say to keep several things in mind:
- The underlying dataset are “tasks” – a set of instructions paired with an output that is the result of the task
- Make sure the task is autogradable via the options we make available in the API (it should be easy to verify if the task was done correctly or not). We currently support a set of graders, but will expand over time. I know this is a bit hard because you can’t see the set of graders available until we launch more broadly – but easily gradable tasks (i.e. string match) are more likely to work out of the box.
- Make sure that the task is clear enough that if expert humans do it, they also converge onto the same answer.
More to come here soon!
The microcontroller was an ESP32-S3.
@michellep Is structured outputs coming to o1mini API soon?