Reference: Models Overview
| Models | Intro | Note |
|---|---|---|
GPT-3.5 (a set): gpt-3.5-turbo is recommended |
understand & generate natural language / code (language model) | gpt-3.5-turbo, gpt-3/5-turbo-0301, text-davinci-003, text-davinci-002, code-davinci-002 |
| DALL-E | generate & edit images given natural language prompt (image generation) | |
| Whisper | convert audio into text (speech recognition) | |
| Embeddings (a set) | convert text into a numerical form (text-relateness measurement) | |
| Codex (a set) | understand & generate code, including translating natural language to code (coding suggestion) | code-davinci-002, code-cushman-001 |
| Moderation | (fine-tuned) detect whether text may be sensitive/unsafe (ie sentiment analysis) | |
| GPT-3 (a set) | understand & generate natural language (language model) | text-curie-001, text-babbage-001, text-ada-001, davinci, curie, babbage, ada |
| Point-E | generate 3D print clouds from prompts (3D-model generation) | |
| JukeBox | generate music from prompt (music composition) | |
| CLIP | generate text snippet from image (caption generation) | |