batuhan Sometimes it's extremely slow and sometimes some models don't generate responses (especially with image prompts) It should be more stable.
Dexter I had similar errors in the past when I asked some recent news and Kagi did not do a websearch with the assistant. Maybe granting webaccess could be a workaround. Of course this situation should be fixes.
batuhan I only uploaded 2 PDFs; it wasn't even 40k tokens. Just apply the model parameters as they are; otherwise, what's the point of having a model that supports 1 million tokens?
Luis Our apologies for these; we've been making quite a few improvements in the background... are you still experiencing issues with the Assistant?
batuhan Luis I'm still having issues with long texts and files. Even though the fair usage limit has been arrived, why is the input/output still below the model's capacity? Even though I paid $25, I still have to go and use ChatGPT and Google AI Studio
