Make Venice Large, well, Large

We are waiting on a bug fix in the underlying server stack. We had expected it this week, but it seems like it’s coming next week. Once that’s out, we should be able to update the context to 128k.

We are exploring large context model options. If there are any specific ones you would like to see, please let us know.

The new large model is a wonderful writing aid but craps out because the 32k window is too small. DeepSeek at 64k also gets stuck but the 405B model, handles it well but is nowhere as good as the new Large model. Medium is all we’ll be left with and it’s just not in the same league as the other two (DeepSeek is nuts though).

I would love to see the Large Model be upgraded to a minimum of 128k with 256k or higher being a creative win for us writers

To add to my comment for a larger 256k model was the need to upload previously worked on stories. When you export them but import them again all the models choke. The siloing off of the threads, which I know is for the privacy, means there has to be a way to give the models the historyical context for developing a story.

It would be nice to have at least 256k context for dev. but honestly if we can get close to the 1024 context of gemini that would be so cool. Dev can be sometimes expensive in context.

I have been waiting for this!!

is there any update on the bug fix and upgrade?

We’re close here, and will be rolling this out as soon as we can.