[🔒 Closed] 4.0 pro empty respopnse from api chub ai

[still cobalt]

CosmosRP 4.0 PRO is supported, but if it gives an empty response on Venus/Chub while 4.0 works, it’s usually a request/config issue, not the model itself.

Try this:

• Use the correct model ID: pkrd/cosmosrp-4.0-pro
• Make sure your request includes a valid API key every time
• Set the base URL to:

  https://api.pawan.krd/v1
  

  or lock it to the model:

  https://api.pawan.krd/pkrd/cosmosrp-4.0-pro/v1
  

• Remove unsupported or conflicting parameters
• Keep temperature in the supported range for this model: 0.4–1
• If Venus has a “thinking” or “reasoning” toggle, try turning it off/on to test

If you want to compare, CosmosRP 4.0 is more permissive on some plans and may behave more reliably depending on your subscription. If the issue continues, it’s best to check your Venus/Chub integration settings since the API side is usually fine when the normal 4.0 model works.

-# This is an automated response from @still cobalt. While I strive to provide accurate assistance, I may occasionally make mistakes. If you find any inaccuracies or need further clarification, feel free to wait for our community helpers to give you further guidance.

[vocal stag]

it really says empty response from API

[still cobalt]

That usually means the request is reaching the API, but the app is not handling the returned content correctly.

Please check:

• Model ID: pkrd/cosmosrp-4.0-pro
• API key: must be a real key from https://pawan.krd/keys
• Base URL: https://api.pawan.krd/v1
• Endpoint: /chat/completions
• Remove any extra fields Venus/Chub adds that the model may not like
• Try a very simple prompt first

If it still says “empty response from API”, it’s most likely an integration issue in Venus/Chub rather than the model. You can also test the same key/model directly with the API to confirm.

-# This is an automated response from @still cobalt. I may occasionally make mistakes — wait for a community helper if you need further clarification.

[vocal stag]

when i add /chat/completion it says : null - The requested endpoint GET /v1/chat/completions/models does not exist

[still cobalt]

You’re using the wrong base URL/path.

Use one of these:

Base URL: https://api.pawan.krd/v1
Endpoint: /chat/completions

or model-locked:

Base URL: https://api.pawan.krd/pkrd/cosmosrp-4.0-pro/v1
Endpoint: /chat/completions

Do not add /models after /chat/completions. That error means the app is building a bad URL like:

/v1/chat/completions/models

Also make sure:

• the API key is valid
• the model ID is exactly pkrd/cosmosrp-4.0-pro

If Venus/Chub has separate fields for base URL and path, put only /chat/completions in the path field.

-# This is an automated response from @still cobalt. I may occasionally make mistakes — wait for a community helper if you need further clarification.

[vocal stag]

ok

[silver reef]

please show us your API settings and your generation settings

[vocal stag]

sure

is that good? xD

[silver reef]

set max new tokens to 2048

[vocal stag]

oh seems like it work xD but how can I tell the bot not to make so big answer? do u have a clue? :X

but still thanks !

[silver reef]

max new tokens forces the model to stop after generating 2048 tokens

its not telling the model to generate short responses, the model still generates the same length response no matter what you set to the max new tokens, its just cuts the response if it gets that length

[vocal stag]

mmmh I see so how can I tell the model to make like one or two paragraph max?

maybe in there?

[silver reef]

you can try to give the model some instructions using the custom prompt, i don't recommend that if you don't know how it works,

the model usually generates the same length as the chat history, you can also try to write short messages

[vocal stag]

ok !

thx again <3 I think we can close

because I think I messed up lmao

[still cobalt]

[🔒 Closed] 4.0 pro empty respopnse from api chub ai