Looks Like ChatGPT API Released!

[Deadlight]

https://openai.com/blog/introducing-chatgpt-and-whisper-apis
https://platform.openai.com/docs/guides/chat

 

[blogmen]

The new model is called gpt-3.5-turbo but doesn't work yet.

Cool thing it's 10x cheaper than davinci.

Anyone managed to call it via API yet?

 

[dvc]

Yup, received an email from them just a couple of minutes ago.

 

[CPANathan]

The new model is called gpt-3.5-turbo but doesn't work yet.

Cool thing it's 10x cheaper than davinci.

Anyone managed to call it via API yet?

I think you mean 10% bro bro!

 

[blogmen]

I think you mean 10% bro bro!

Nope. Text-davinci-003 costs $0.02/1k tokens and the new model $0.002/1k tokens and has the same output quality lol.

 

[CPANathan]

Nope. Text-davinci-003 costs $0.02/1k tokens and the new model $0.002/1k tokens and has the same output quality lol.

Oh wow, I read the email and thought even 10% cheaper would be amazing...

Mind is blown, thanks for correcting me and not being cheesy

 

[aperfectcircle]

This is getting better X D

 

[RRX]

Nope. Text-davinci-003 costs $0.02/1k tokens and the new model $0.002/1k tokens and has the same output quality lol.

they got two of them

 

[anonsops]

so we move from davinci-003 to gpt-3.5-turbo

 

[Lukmat]

So, the era of autoblogs begin xD

 

[710fla]

The new model is called gpt-3.5-turbo but doesn't work yet.

Cool thing it's 10x cheaper than davinci.

Anyone managed to call it via API yet?

That's crazy considering davinci is dirt cheap alright

 

[Daemon Targaryen]

Too cheap. Looks like all the free api keys that Sartre gave away will be used on this model.

 

[Deadlight]

Too cheap. Looks like all the free api keys that Sartre gave away will be used on this model.

The timing couldn't have been better

 

[wolfman1]

This is the bit where our hero Google, has taken a sword to the stomach … and the villain (OpenAI) walks up, smirks to the crowd … and violently rips the sword back out, exposing a giant, fatal gash …

Good night sweet prince Google … your time is over.

 

[NemoTheOne]

So, the era of autoblogs begin xD

You mean, the era returns!

 

[vpnsaresmart]

Does the API suffer the same speed limitations?

 

[V]

Too cheap. Looks like all the free api keys that Sartre gave away will be used on this model.

I hope I also get some API keys from Sartre. I've been procrastinating it long enough.

 

[blincusa]

I really look forward to using an API to create services for ChatGPT

 

[Dj Co2]

View attachment 244934
they got two of them

Why am I not getting them in my playground?

 

[maresen]

The new model is called gpt-3.5-turbo but doesn't work yet.

Cool thing it's 10x cheaper than davinci.

Anyone managed to call it via API yet?

that works:

import os
import openai
openai.api_key = "sk-gggggggggggggggggggggggggggggggggggg"


response = openai.ChatCompletion.create(
  model="gpt-3.5-turbo",
  messages=[
    {"role": "system", "content" : "you are an expert in ?????  "  },
    {"role": "user", "content": " provide ?????????? "},
    ],
   temperature=0.3,
)
print(response["choices"][0]["message"]["content"])