OpenAI Slashes its Prices by 50%

Recently, OpenAI announced that they are slashing their prices by more than 50%.

https://twitter.com/deliprao/status/1561783712152428544

The new pricing is set to take effect from 1 September, 2022. This would make the base models more affordable, with support for creating custom models by fine-tuning.

Decreased prices would also allow users to access ‘Ada’, the fastest API model along with ‘DaVinci’, the most powerful API.

Users can begin for free with $18 credit for the first three months. Upon signing up, they would be granted an initial spend limit—or quota—to make the rollouts easier. The limit would increase as users build a track record with the application.

Tokens for creating a custom model by modifying the base model are billed at 50% of the base price. Users can also build advanced search, clustering, topic modeling and classification functionality by embedding models.

Users can interact with the API through HTTP requests from languages like Python for which the search and answer requests are billed on the numbers of tokens input by the user along with the tokens needed to instruct the model on the operation.

Prices are per 1000 tokens, where one token is approximately four characters or three-fourth of a word. 

Developers using the OpenAI API with the Moderation Endpoint would now have free access to GPT-based classifiers that detect objectionable content faster and with greater precision.

The price drop of OpenAI makes it more affordable than CohereAI but it is still twice as expensive as GooseAI.

OpenAI API is also compatible with Microsoft Azure in terms of compliance, enterprise grade security, and regional support.