Hi all. Question about GPT-4 access. I see on the help center: API accounts created after August 18, 2023, you can get instant access to GPT-4 after purchasing $0.50 worth or more of pre-paid credits (see screenshot)
I just created a new API account and deposited $10, but don’t have GPT-4 API.
Any idea what is going wrong here?
Thanks
The next step in the troubleshooting process would be to generate a new API key.
Then use that key to call the models endpoint for an account model listing. A python code quickie for printing just their names:
import openai
openai.api_key = "sk-xxx"
try:
models = openai.Model.list()
model_ids = [model.id for model in models["data"]]
print(model_ids)
except Exception as err:
print(f"Error updating model list: {err}")
There has been more reported success of instant GPT-4 access when enrolling in the new prepaid credit system than any mentions of problems (except for problems originating from limitations of the plan service itself)
Hi. I am facing the same problem! I’ve just ran the code bit you provided and I do not have there GPT-4 in the list.
Also the GPT-4 should be visible in the playground, is this correct? If so, I don’t haver this also visible there. What other solutions here?
Thanks
Ok! got it working … but needed to go for the Pay As You Go option, there is no way to get it as a rolling service?
New API users that are only now putting a payment method into the API billing interface are only given the option to purchase a prepaid credit, minimum $5. Not getting gpt-4 instantly after doing so would be a bug.
Prior accounts that already input a credit card have post-paid billing. GPT-4 access was given a month ago to those who had done so and also incurred billings against their payment method.
The prepaid help documentation describes in the future they may allow post-paid conversions, larger payments, auto-refills.
Thanks so much! This worked and can confirm gpt 4 api access.
A month or so ago I signed up for API access (before was only using the chat interface). I’m doing the “Pay as you go” billing. Then I created an API key and tried to use it, but was given an error when trying to use it:
The model 'gpt-4' does not exist or you do not have access to it
I tried using your code, after doing a
❯ pip install openai
# openai.py
import openai
openai.api_key = "sk-...."
try:
models = openai.Model.list()
model_ids = [model.id for model in models["data"]]
print(model_ids)
except Exception as err:
print(f"Error updating model list: {err}")
But got this error:
❯ python openai.py
Error updating model list: partially initialized module 'openai' has no attribute 'Model' (most likely due to a circular import)
Error updating model list: module 'openai' has no attribute 'Model'
Any chance I can get access to GPT-4?
Openai python library requirements:
Python version 3.7.1-3.9 (3.10 should work for most calls)
pip install --upgrade openai
If python was installed “For all users” in Windows, then administrator cmd.exe or powershell must run pip to put packages in the system install.
I just pasted the model list code from the forum into a text editor, added API key (generate new to ensure current models), ran, results:
['davinci', 'text-similarity-curie-001', 'gpt-3.5-turbo-16k-0613', 'gpt-4-0314', 'text-davinci-001', 'babbage', '...
The playground will also show it.
For a monthly billing pay-as-you-go account, you must incur billed use to be considered for gpt-4. You’ve got another 15 days to use the API to the tune of $1+ after trial, and another 15 days to get charged.
Opening a second account for prepaid billing and feeding it money will get that account gpt-4 in the interim.
When I run this command I see the following:
❯ pip install --upgrade openai
Requirement already satisfied: openai in ./.venv/lib/python3.10/site-packages (0.27.10)
Collecting openai
Downloading openai-0.28.0-py3-none-any.whl (76 kB)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 76.5/76.5 kB 4.0 MB/s eta 0:00:00
Requirement already satisfied: aiohttp in ./.venv/lib/python3.10/site-packages (from openai) (3.8.5)
Requirement already satisfied: tqdm in ./.venv/lib/python3.10/site-packages (from openai) (4.66.1)
Requirement already satisfied: requests>=2.20 in ./.venv/lib/python3.10/site-packages (from openai) (2.31.0)
Requirement already satisfied: certifi>=2017.4.17 in ./.venv/lib/python3.10/site-packages (from requests>=2.20->openai) (2023.7.22)
Requirement already satisfied: charset-normalizer<4,>=2 in ./.venv/lib/python3.10/site-packages (from requests>=2.20->openai) (3.2.0)
Requirement already satisfied: urllib3<3,>=1.21.1 in ./.venv/lib/python3.10/site-packages (from requests>=2.20->openai) (2.0.4)
Requirement already satisfied: idna<4,>=2.5 in ./.venv/lib/python3.10/site-packages (from requests>=2.20->openai) (3.4)
Requirement already satisfied: async-timeout<5.0,>=4.0.0a3 in ./.venv/lib/python3.10/site-packages (from aiohttp->openai) (4.0.3)
Requirement already satisfied: multidict<7.0,>=4.5 in ./.venv/lib/python3.10/site-packages (from aiohttp->openai) (6.0.4)
Requirement already satisfied: aiosignal>=1.1.2 in ./.venv/lib/python3.10/site-packages (from aiohttp->openai) (1.3.1)
Requirement already satisfied: attrs>=17.3.0 in ./.venv/lib/python3.10/site-packages (from aiohttp->openai) (23.1.0)
Requirement already satisfied: yarl<2.0,>=1.0 in ./.venv/lib/python3.10/site-packages (from aiohttp->openai) (1.9.2)
Requirement already satisfied: frozenlist>=1.1.1 in ./.venv/lib/python3.10/site-packages (from aiohttp->openai) (1.4.0)
Installing collected packages: openai
Attempting uninstall: openai
Found existing installation: openai 0.27.10
Uninstalling openai-0.27.10:
Successfully uninstalled openai-0.27.10
ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
open-interpreter 0.1.4 requires openai<0.28.0,>=0.27.8, but you have openai 0.28.0 which is incompatible.
litellm 0.1.610 requires openai<0.28.0,>=0.27.8, but you have openai 0.28.0 which is incompatible.
For a monthly billing pay-as-you-go account, you must incur billed use to be considered for gpt-4. You’ve got another 15 days to use the API to the tune of $1+ after trial, and another 15 days to get charged.
Opening a second account for prepaid billing and feeding it money will get that account gpt-4 in the interim.
I’m in a loop, I cannot use gpt-4 without bill use, but how do I incur billed use when I cannot use gpt-4?
Also, I don’t think it matters, but I’m a Plus user, so I use GPT-4 chat all the time.
There’s lots of other models to use.
Send gpt-3.5-turbo-16k 15000 tokens, 20 times = $1.20.
You can probably find more practical API uses to prepare your skills.
Thanks for your response. Can I use the playground to make the requests?
Will this incur the billing?
Yes, you get billed for playground use just as if you were using the API. It is essentially accessing the API for you, automatically taking care of account API keys.
Normal chats are quite inexpensive, however the costs of a continuing conversation can grow, as all that you see is sent every time you press submit, in order to demonstrate AI understanding from past conversation turns.
Run this and be a dollar poorer in a minute:
import openai
import time
openai.api_key = "sk-xxxx"
params = {"model": "gpt-3.5-turbo-16k", "max_tokens": 1,
"messages": [{"role": "user",
"content": "!@" * 8000}]}
for _ in range(21): # 0.003 * 16 * 21
api = openai.ChatCompletion.create(**params)
print(".", end="")
time.sleep(0.2)
Thanks but with this code, I see this error message:
❯ python openai2.py
Error updating model list: partially initialized module 'openai' has no attribute 'Model' (most likely due to a circular import)
Traceback (most recent call last):
File "/Users/aaron/dir/openai2.py", line 8, in <module>
api = openai.ChatCompletion.create(**params)
AttributeError: module 'openai' has no attribute 'ChatCompletion'
I ran this:
❯ pip install --upgrade openai
Requirement already satisfied: openai in ./.venv/lib/python3.10/site-packages (0.28.0)
Requirement already satisfied: aiohttp in ./.venv/lib/python3.10/site-packages (from openai) (3.8.5)
Requirement already satisfied: requests>=2.20 in ./.venv/lib/python3.10/site-packages (from openai) (2.31.0)
Requirement already satisfied: tqdm in ./.venv/lib/python3.10/site-packages (from openai) (4.66.1)
Requirement already satisfied: urllib3<3,>=1.21.1 in ./.venv/lib/python3.10/site-packages (from requests>=2.20->openai) (2.0.4)
Requirement already satisfied: certifi>=2017.4.17 in ./.venv/lib/python3.10/site-packages (from requests>=2.20->openai) (2023.7.22)
Requirement already satisfied: charset-normalizer<4,>=2 in ./.venv/lib/python3.10/site-packages (from requests>=2.20->openai) (3.2.0)
Requirement already satisfied: idna<4,>=2.5 in ./.venv/lib/python3.10/site-packages (from requests>=2.20->openai) (3.4)
Requirement already satisfied: attrs>=17.3.0 in ./.venv/lib/python3.10/site-packages (from aiohttp->openai) (23.1.0)
Requirement already satisfied: yarl<2.0,>=1.0 in ./.venv/lib/python3.10/site-packages (from aiohttp->openai) (1.9.2)
Requirement already satisfied: frozenlist>=1.1.1 in ./.venv/lib/python3.10/site-packages (from aiohttp->openai) (1.4.0)
Requirement already satisfied: aiosignal>=1.1.2 in ./.venv/lib/python3.10/site-packages (from aiohttp->openai) (1.3.1)
Requirement already satisfied: async-timeout<5.0,>=4.0.0a3 in ./.venv/lib/python3.10/site-packages (from aiohttp->openai) (4.0.3)
Requirement already satisfied: multidict<7.0,>=4.5 in ./.venv/lib/python3.10/site-packages (from aiohttp->openai) (6.0.4)
[notice] A new release of pip is available: 23.0.1 -> 23.2.1
[notice] To update, run: pip install --upgrade pip
Don’t name your script openai or similar, or have one in your path.
reduce 8000 (= 16000 tokens) to something like 8 to experiment with the looping call.
Great, so I reached a dollar:
Current usage
Your total usage so far in September (UTC). Note that this may include usage covered by a free trial or other credits, so your monthly bill might be less than the value shown here. View usage records
$1.77
So what do I need to do in order to get GPT-4 as an option now?
I ran the list of models and this is what I see:
['davinci', 'text-similarity-curie-001', 'gpt-3.5-turbo-16k-0613', 'text-davinci-001', 'babbage', 'text-babbage-001', 'curie-similarity', 'gpt-3.5-turbo-16k', 'curie-instruct-beta', 'code-davinci-edit-001', 'davinci-search-query', 'code-search-ada-code-001', 'text-curie-001', 'davinci-search-document', 'ada-code-search-text', 'davinci-002', 'gpt-3.5-turbo-0613', 'text-search-davinci-query-001', 'code-search-ada-text-001', 'ada-code-search-code', 'babbage-002', 'babbage-search-document', 'text-search-curie-query-001', 'text-search-davinci-doc-001', 'text-embedding-ada-002', 'gpt-3.5-turbo', 'davinci-instruct-beta', 'ada-search-query', 'text-search-babbage-query-001', 'text-search-curie-doc-001', 'text-davinci-002', 'babbage-code-search-code', 'ada-search-document', 'text-search-babbage-doc-001', 'whisper-1', 'text-davinci-edit-001', 'ada', 'davinci-similarity', 'code-search-babbage-code-001', 'text-ada-001', 'text-similarity-davinci-001', 'code-search-babbage-text-001', 'gpt-3.5-turbo-0301', 'curie', 'babbage-code-search-text', 'text-search-ada-query-001', 'text-similarity-babbage-001', 'curie-search-document', 'text-similarity-ada-001', 'text-search-ada-doc-001', 'babbage-similarity', 'ada-similarity', 'babbage-search-query', 'text-davinci-003', 'curie-search-query']
Get a September bill in October. If they still have the same onboarding policy in place, you should get access after you are charged.
So I have to wait until the next billing cycle?
Yes, as I wrote in the reply to you above three hours ago:
Sorry for being annoying.
Thanks for everything.