[Question] What will it take to get the 4o image gen working in LibreChat?

[dejer]

Is there a known way to do this already? I'd rather use API access instead of paying $20/month for GPT Plus.

Is there a specific model that should be used for 4o? Is "chatgpt-4o-latest" the best, or just "gpt-4o"? Something else entirely? Do I need to use an agent for this, or can it handle an image-to-image request natively?

Lastly, does this need to go through an Agent Builder with DALL-E-3? It doesn't seem like that is the right approach.

[Note: LibreChat Local Docker install updated as of 3/28/2025]

[TonyCollett]

Last time I looked, it wasn't available through the API yet. That was a few days ago though. I think the OpenAI website said 'coming to API in a few weeks' or something along those lines.

[danny-avila]

Yes it’s not out yet, track the open issue: #6592