agentic

agentic copied to clipboard

Verify latest release

• [X] I verified that the issue exists in the latest chatgpt release

Verify webapp is working

• [X] I verify that the ChatGPT webapp is working properly for this account.

Environment details

18.0.0

Describe the Bug

Hi, results are different between api and browser

thx

Maybe they using defferent model

Pretty sure the new version of this package is using davinci, not chatgpt models which is what runs in the browser, so wouldn't expect the same results.

Pretty sure the new version of this package is using davinci, not chatgpt models which is what runs in the browser, so wouldn't expect the same results.

can customize model?

Pretty sure the new version of this package is using davinci, not chatgpt models which is what runs in the browser, so wouldn't expect the same results.

can customize model?

Yes @kknono, here’s how to override the default model

const api = new ChatGPTAPI({
    apiKey: process.env.OPENAI_API_KEY,
    completionParams: {
      model: 'text-davinci-003'
    }
})

This is unfortunately expected for the time being. See my detailed answer here https://github.com/transitive-bullshit/chatgpt-api/issues/314#issuecomment-1422395492

thx

text-chat-davinci-002-sh-alpha-aoruigiofdj83 available now~

@SmileSmith this model 404's for me. Where did you find it?

Nevermind; it got disabled.

Why is this lib misleading developers advertising the package as "ChatGPT" when it actually uses text-davinci-003? model should be text-chat-xxx...

@telemakhos see my previous answer here https://github.com/transitive-bullshit/chatgpt-api/issues/314#issuecomment-1422395492

You can now use ChatGPTUnofficialProxyAPI which uses the a proxied version of the real ChatGPT API. Upgrade to https://github.com/transitive-bullshit/chatgpt-api/releases/tag/v4.5.0 and check out the readme for usage info and notes on tradeoffs.

Browser version use "gpt-3.5-turbo" model.

Usage example (Python):

openai.ChatCompletion.create(
 model="gpt-3.5-turbo",
 messages=[
       {"role": "system", "content": "You are a helpful assistant."},
       {"role": "user", "content": "Who won the world series in 2020?"},
       {"role": "assistant", "content": "The Los Angeles Dodgers won the World Series in 2020."},
       {"role": "user", "content": "Where was it played?"}
   ]
)

More details here - OpenAI Documentation

how can I override the default modeland use "gpt-turbo"model?I mean which file should I edit? And should I run any thing after edit? thank you very much!

how can I override the default modeland use "gpt-turbo"model?I mean which file should I edit? And should I run any thing after edit? thank you very much!

Here is snippet from repository READMY.md:

const api = new ChatGPTAPI({
  apiKey: process.env.OPENAI_API_KEY,
  completionParams: {
    temperature: 0.5,
    top_p: 0.8
  }
})

It looks pretty similar to my Python example, so you can try to add model="gpt-3.5-turbo" parameter, maybe it will work (I'm Python developer, that just make research about difference between browser and api versions)

@BobWangRobot @AndrejGorodnij just update to the latest v5 of this package and use ChatGPTAPI, which already uses the gpt-3.5-turbo model.

ok, thank you very much!