Several variations of the GPT-4 model exist, each distinguished by unique features.
A crucial attribute that varies between these models is the context length, which denotes the total number of tokens consisting of the prompt and the maximum tokens in the completion. The standard GPT-4 model provides a context length of 8,000 tokens. Additionally, OpenAI introducing an upgraded model that allows a context length of 32,000 tokens, which is being launched independently of the 8k version.
Furthermore, the term gpt-4 refers to the most recent model version, with gpt-4-32k indicating the latest 32k-context model. If you wish to use a previous model iteration, you can do so by specifying the date in the model’s name, such as gpt-4-0314 or gpt-4-32k-0314 for the March 14th version. Keep in mind that the March 14th version will remain accessible until June 14th.
How can I access GPT-4?
API Access
For API access, many users will need to be placed on waitlist. OpenAI striving to provide access to all as quickly as possible, however, it might take some time before everyone gains access to GPT-4. Developers sincerely thank you for your patience and eagerly anticipate the innovative applications you’ll create with the API!
ChatGPT Plus
Subscribers to ChatGPT Plus will gain access to GPT-4 on chat.openai.com, albeit with a usage limit. The exact cap on usage will be subject to periodic adjustments based on the demand and system performance. Currently, with GPT-4, the limit is set at 25 messages every 3 hours.
You also can find more information about OpenAI API cost on Pricing page.