Using GPT-3 API can help you save time and money by automating tasks such as natural language processing, text classification, and search. The cost of GPT-3 API is based on the number of tokens in the inputs you provide and the models you select. The cost per token is calculated based on the search and completion models you select and the length of the query passed into the model. You can also provide a file containing examples to search over or explicitly specify examples in your request. With GPT-3 API, you can reduce costs by providing files with more examples and reranking them. Get ready to save time and money with GPT-3 API!
Language Models
| Name | Price/1K tokens |
|---|---|
| Ada | $0.0004 |
| Babbage | $0.0005 |
| Curie | $0.0020 |
| Davinci | $0.0200 |
There are multiple models that each offer different capabilities, are trained for different specialists, and come at different price points. The Ada model is the fastest of the group, while the Davinci model is the most powerful. Those who need speed should consider the Ada model, while those looking for greater power may opt for the Davinci. All models come with varying features and advantages, so regardless of which one you choose, you will be getting an optimised a.i.
The prices for OpenAis services are based on a 1,000 tokens system. To put it in simpler terms, each 1,000 tokens is equivalent to approximately 750 words. For example, this sentence you’re reading right now is around 30 tokens.
Fine Tuned Models
You can create custom models specifically tailored to your data by fine-tuning our base models. After you have fine-tuned a model, you’ll only be charged for the amount of tokens that you use when making requests to that model. This is an excellent way to ensure that you are getting the most out of your models while keeping costs low.
| MODEL | TRAINING | USAGE |
|---|---|---|
| Ada | $0.0004 / 1K tokens | $0.0016 / 1K tokens |
| Babbage | $0.0006 / 1K tokens | $0.0024 / 1K tokens |
| Curie | $0.0030 / 1K tokens | $0.0120 / 1K tokens |
| Davinci | $0.0300 / 1K tokens | $0.1200 / 1K tokens |
To get the most out of your fine-tuned models, it’s important to understand the two components of pricing: training and usage. When training your model, you’ll be charged for the total number of tokens used, which is based on the number of tokens in your training dataset and the number of training epochs you choose. By default, training epochs are set to 4. After your model is fine-tuned, you’ll only be charged for the tokens you use when making requests. We’re here to help you make the most of your fine-tuned models, so if you have any questions about pricing, please don’t hesitate to reach out.
Embedding models
| MODEL | USAGE |
|---|---|
| Ada | $0.0004 / 1K tokens |
| Bert | $0.0025 / 1K tokens |
| GPT-3 | $3.00 / 1K tokens |
You can utilize their embeddings offering to develop sophisticated search, clustering, topic modeling, and classification features. With these features, you will be able to search for specific terms, group similar terms together, explore, and classify content more effectively.
Image models
| RESOLUTION | PRICE |
|---|---|
| 1024×1024 | $0.020 / image |
| 512×512 | $0.018 / image |
| 256×256 | $0.016 / image |
| 128×128 | $0.014 / image |
You can take advantage of the power of DALL·E by integrating it directly into your apps. With three tiers of resolution, you can create and edit unique images and artwork with a variety of sizes. The image generations endpoint allows you to use a text prompt to create an entirely original image. Requesting 1-10 images at a time, you can choose from sizes of 256×256, 512×512, and 1024×1024 pixels. Smaller sizes are faster to generate, so you can bring your ideas to life quickly.
You can increase your chances of getting the desired result by making sure your description is as detailed as possible. To get some ideas, you can check out the examples in the DALL·E preview app. It’s a great way to get your creative juices flowing!
