The grok-beta, grok-2-1212, gemini-2.0-flash-exp models are now available to premium users only

Today, there may be errors as the transfer to another server will be taking place

At the moment, the api can work very badly, very soon everything will be fixed🙏

It's working fine now, I'll try not to let it happen again

🔥 New update

⚡️ New models o3 (premium), o4-mini (premium) added

🔥 Small update

⚡️ A lot of open-source models have been removed (check at api.gpt4-all.xyz/v1/models). Also image generation has been permanently removed

🔥 New update
Deleted models gemini-1.5-pro, gemini-1.5-flash, gemini-1.5-flash-8b, gemini-2.0-flash-exp, gpt-3.5-turbo
Added new models gemini-2.5-pro, gemini-2.5-flash
And API now is more stable

🔥 Small update
Added new model kimi-k2-instruct with 1 trillion total parameters
Its smart, cheep and fast!
Best LLM from Chinese language

GPT-5 still thinks he is GPT-4🫢
I dont know when GPT-5 will be added

🔥 Small change
Default value for max_tokens is now 1024
If you got an empty response from the thinking model, just increase max_tokens. Think tokens are also counted

error with this model has been fixed, sorry!

default limit on max_tokens was deleted, for some reasons, i would ask you dont spend tokens too many, just add in system prompt "answer briefly without unnecessary words", or something like this