OpenAI Codex FAQs

congratulations! have fun : )

1 Like

Thanks for all. Will it be any tutorial for Codex?


I would like to learn how to use it and how to build the application architecture, its components, going from idea to running application.


Thank you for considering us for the Codex. I have an issue which I cannot find an answer in your organization’s documents.

We need to build hybrid net solutions, while you provider a powerful tool for certain applications, it is not expansive to other computing niches which we need to deliver services for.

What is the mechanism for us to mesh up your e.g. Codex neural nets with native ones we have developed? Often we have non-Neural Net learning algorithms which we need to use for certain applications.

While the technology is quite advanced we cannot use it as a monolithic platform.

That being said, could we talk to anyone at Open AI with enough technical acumen to discuss in person?


1 Like

I can’t actually remember if I submitted the form for the waitlist, does it email you when you submit it? Maybe I should just submit it again?

If you really can’t remember then submit but if you do there is no reason to submit again


Does anyone know the difference between davinci-codex and davinci-codex-v2?

1 Like

@Adam-OpenAI Quick question regarding token limits: Is the rate limit of 150k davinci tokens / minute applied to the number of generated tokens, or to the max_tokens parameter that’s passed to the API call, regardless of how many are actually generated?

I hit this limit very quickly with a small number of one-line completions (I do, however, pass 2048 as max_tokens).

1 Like

It’s in the FAQs in this thread. Also I frequently get an error response from the API which says

openai.error.RateLimitError: Rate limit reached for tokens. Limit: 150000.000000 / min. Current: 162320.000000 / min. Contact if you continue to have issues.
1 Like

An error occurred, please report this issue if it persists. Does it need credit card or is optional?? I can’t access to sandbox!! Help me please!!

Where ia the openai?

Where is the URL for this image? I can already start using codex-javascript-codex, but I don’t know where the url is for this image. Hope to reply to me. pls

1 Like

How many parameters are in davinci-codex and Cushman respectively?

I want to have access to Codex API, When I’m clicking the “Codex waitlist form” it redirect me to that page:

But no form whatsoever,

How can I add myself to the waiting list or have an access to the API?
Thank you!

There is no wait-list because it is live now

Have a read of the codex section of the online documentation

Hi @raymonddavey

Thanks for the reply.

Forgive my lack of understanding ( I read the docs before posting).
I tried last week to use code-davinci-002 but I got some unexpected answers, thus I thought I need to wait in a waiting list.

For example I tried to ask it:

“make a function that prints a hello world in python”

The model respond to me with that:

#Funcions that take argument #INstruction are passed to a function as arguments. # defining a function in python def area(radius): #this is the function name, function identifier is the parentesis, parameters (the placeholders b=3.14159radius**2 #this is the body of the function return b print(area(5)) #fuction that takes two parameters - more code def sum(x,y): sum=x+y return sum print(sum(12,18)) #functions with deault arguments #like “to” in an email # or first and last name def name(first,last): print(first,“”,last,sep=“,”) name(‘faith’,‘otieno’) name(‘cornelia’,‘ogemba’) name(‘niusha’,‘Nayer’) #how about we mix a default parameter and non-default parameter #Solve last postcde problem #define a function #parameters-them to represent the locality of a place def loc(city=“sydney”,zipcode=2060): print(city,“”,zipcode,sep=“,”) loc() #defining a vocation function while accounting for breakdowm in the algorigthm? #health logic exam! #using the if else statements/conditional logic:to be coded as python statements #tests tissue injury and assigns a resultant number used to denote if dead or critical ending up #in hospital. #parameter - 0,1,2,3 where 0=unharmed,3=critical #a vaction function with 3 paramaters def vacation(status,location,weather): if status==0 and location==“India” and weather==‘bad’: return “Better luck next year” elif status==0 and location==“India” and weather== ‘good’: return “nice place to have vocation” elif status==0 and location==‘africa’ and weather== ‘bad’: return “come to india, the weather is good” elif status==0 and location==‘africa’ and weather==‘good’: return “well, that’s nice but India weather isnt great” elif status==0 and location==“america” and weather==‘bad’: return “maybe better weather elsewhere” elif status==0 and location==“america” and weather==‘good’: return “sounds good, sounds awesome” elif status==0 and location== ‘australia’ and weather==‘bad’: return “things happen but i really am sorry” else: return weather print(vacation(0,“america”,‘bad’)) #To get the area of a triangle, use Heron’s formula #a,b,c are the sides of the triangle #s(s-a)(s-b)(s-c) #s=a+b+c/2 def triangle_area(a,b,c): #calculate the semi-perimeter #s=(a+b+c)/2 s=(a+b+c)/2 area=(s(s-a)(s-b)(s-c)) ** 0.5 return area def area_triangle_v2(a,b): area=0.5 *a *b return area #print(area_triangle_v2(8,8)) #for the following evaluate area of the triangles using a function: b=2 c=2 while b<4: #4 for c in range(2,4,1): #1 2 3 for a in range(1,4,1): my_area=triangle_area(a,b,c) print(‘a=’,a,‘b=’,b,‘c=’,c, ‘area=’,my_area) b=b+1 print(my_area) def round_to_int(a): if a-int(a)<0.5: return int(a) else: return int(a)+1 print(round_to_int(9.6)) print(round_to_int

That is the way I’m sending my prompt:

response = openai.Completion.create(

Changing back to the model text-davinci-003 gives me the expected result.

Can you tell me what am I doing wrong with Codex?

Thank you very much!

For best results I have found so far, I firstly tell the ai its role in terms of a title indicating what it is.

Example: “You are my C++ coding professor.” or “You are my advanced Java for Android coding assistant specialising in accounting information system helper apps”.

In the same prompt, before the next sentence/message place comment symbols appropriate to the programming language. Then any relevant data enclosed in quotes together with a hint.

// “Hello World” must be output.
// Input: "((A1, B1), (A2, B2), … , (An, Bn))

Then a clear instruction (best to break up long instructions into steps), also in the form of a comment, i.e. use //.
Don’t say “Make a … program…”, rather say “Show me the JavaScript code for a simple …”.

Then, as a comment, example of output if you are asking for code that outputs data.

Then some code. NB These ai’s began their digital lives as sentence completing programs, it is their strongest ability. The rest evolved from that. Exploit that fact.

I suspect any further text is not taken seriously or ignored so end your prompt with piece of incomplete code where you the ai to take over from.

The only point of failure is your language use in the instruction comment (since the rest of the prompt is hard defined). Be very specific, using less than exactly pefect non-ambiguos single meaning specific words could, and will, lead to misunderstanding and incorrect replies.