We just swapped in a new image model, FLUX. Many users consider it superior. I encourage you all to try it out and give us feedback.
6 Likes
images were working ok with dalle, why not give us the option of flux if we choose. my experience is dalle 3 is better.
here is what I got for asking for an image just now: Invalid prisma.studioUser.findFirst() invocation: Raw query failed. Code: unknown. Message: `Kind: Server selection timeout: No available servers. Topology: { Type: Sharded, Servers: [ { Address: ac-5klm6qh-shard-00-00.io3ldgy.mongodb.net:27016, Type: Unknown, Error: Kind: I/O error: failed to lookup address information: Temporary failure in name resolution, labels: {} }, { Address: ac-5klm6qh-shard-00-
DALLE-3 was giving more problems to more users after some content filter updates. Many users were finding the filter both draconian and temperamental. FLUX seems to be a superior model and quite good at images with lettering.
We’re monitoring errors with it pretty closely and within a few days all the errors and kinks should be worked out!
again, why not give users the option to choose which model they want to use. have both available.
Awesome! Results are definitively superior and better quality with Flux. Cool!
May I know which version it is (I’m away and can’t connect my laptop right now 
)
Now you can. Check out the “act” tab in the builder.
I like FLUX schnell better
1 Like
You’re in luck, that’s the version we’re running
you think dalle is better ?
I’ve never heard anyone say DallE was better for anything.
Niceeee. Will you add the other versions of Flux in the future?
Probably. Through our new actions building feature, we will probably enable a ton of different image models.
including dalle again ? flux might be better for geeks but for simple prompts like I use I got great results from dalle and NOT with flux. Thinking of refunding due to this
I’d recommend keeping DALL·E as a option at least, especially considering potential future updates from OpenAI.
Also with Flux, the Schnell model Black Forest ranks as their lowest quality. Comparatively, the images produced are subpar, even falling short of DALL·E 3’s standards. For a better user experience with Flux, the Pro or Dev models is strongly recommended.
Hi, how can I go back to using the Dall-E 3 model with my Pickaxe credits? I very strongly prefer the Dall-E model for my images, and it is the reason I bought Pickaxe. I don’t want to use an API key for image generation.
Thanks,
Mike
You still can use Dall E as images generation model by activate this ACT on Dalle openai API.
But, in the prompt must insert ‘Generate using DALLE 3’ to trigger your bot to use DALLE instead of default Flux model.
I’m doing this way, and its works super great.
2 Likes
You can still use DALL-E3. Just go to Actions and use the DALLE action. It will require you to enter your OpenAI API key.
The beauty of our new Actions system it will allow us to connect many images models over time very simply.
I want to use my credits, not my API key. How can I continue to do this?
Thanks,
Mike
You can use our native image generation capability (which is FLUX) for free.
But the DALL-E integration is now an Action. So you have to ‘connect’ it by entering your OpenAI API key.