OpenAI Models List (March 2023)

OpenAI Models List

OpenAI’s API is fueled by a diverse set of models. Each possessing unique capabilities and price points. With fine-tuning, it’s possible to make limited customizations to the original base models that cater to your specific use cases. Below you can find OpenAI Models List available on March 2023.

GPT-4A set of models that improve on GPT-3.5 and can understand as well as generate natural language or code
GPT-3.5A set of models that improve on GPT-3 and can understand as well as generate natural language or code
GPT-3A set of models that can understand and generate natural language
DALL·EA model that can generate and edit images given a natural language prompt
WhisperA model that can convert audio into text
EmbeddingsA set of models that can convert text into a numerical form
CodexA set of models that can understand and generate code, including translating natural language to code
ModerationA fine-tuned model that can detect whether text may be sensitive or unsafe

Also, OpenAI recently published open source models including Point-EWhisperJukebox, and CLIP.

Below you can find short description of each Model.


GPT-4 is a large multimodal model (accepting text inputs and emitting text outputs today, with image inputs coming in the future) that can solve difficult problems with greater accuracy than any of our previous models, thanks to its broader general knowledge and advanced reasoning capabilities. Like gpt-3.5-turbo, GPT-4 is optimized for chat but works well for traditional completions tasks.


The GPT-3.5 models are capable of comprehending and producing natural language or code. Among these models, the gpt-3.5-turbo is the most powerful and cost-effective option. The model is specifically designed for chat-related tasks, but it also performs remarkably well in conventional completion tasks.


The GPT-3 models possess the ability to comprehend and produce natural language. Although they were superseded by the more potent GPT-3.5 generation models, the original GPT-3 base models, including davinci, curie, ada, and babbage, remain the only models available for fine-tuning currently.


DALL·E is an AI system capable of generating realistic images and artwork based on natural language descriptions. With the DALL·E API, you can create new images of a specified size, edit existing images, or generate variations of user-provided images.

API currently supports the second iteration of the DALL·E model. The model generates images with greater resolution, accuracy, and realism than the original model. You can also read an article about comparison Dalle vs Stable Diffusion.


The OpenAI API offers the Whisper model, which is a multi-task speech recognition model trained on a diverse set of audio data. It can perform speech recognition, language identification, and speech translation across multiple languages.

The Whisper v2-large model, also known as whisper-1, is currently available through our API.

Also, OpenAI recently introduced ChatGPT Whisper APIs.


The text-embedding-ada-002 is the second generation embedding model that offers a more cost-effective solution compared to the previous 16 first-generation embedding models.

Embeddings are numerical representations of text that measure the relationship between two pieces of text. They are beneficial in various tasks such as search, clustering, recommendations, anomaly detection, and classification.


The Codex models have the ability to comprehend and generate code. Their training data comprises of natural language as well as billions of lines of open-source code obtained from GitHub.

These models exhibit their greatest proficiency in Python while being skilled in twelve languages such as

  • JavaScript,
  • Go,
  • Perl,
  • PHP,
  • Ruby,
  • Swift,
  • TypeScript,
  • SQL,
  • Shell.

During the limited beta the Codex models are available for free usage but with reduced rate limits. As we gather more information on their usage, OpenAI plan to offer pricing options to allow for a wider range of applications.


The Moderation models aim to verify whether content conforms to OpenAI’s usage policies. These models offer classification abilities to detect content in various categories, such as hate, hate/threatening, self-harm, sexual, sexual/minors, violence, and violence/graphic.