AI provider

Meta

Meta develops Muse, Llama and the Segment Anything Model family. Muse includes managed APIs and separate weight releases. Llama provides downloadable language and vision models. SAM models solve segmentation or reconstruction tasks rather than acting as general chat assistants.

Models & tools

  • A multimodal text-output model with a 1,048,576-token context and selectable reasoning effort.

    Launch date not confirmedAPI · Provider product
  • A multimodal mixture-of-experts model with 816B total and 42B active parameters that returns text.

    Launch date not confirmedDownloadable
  • The smallest Llama 3.1 instruction model, with 8B parameters.

    Launch date not confirmedDownloadable
  • A 70B text-only instruction model for multilingual dialogue.

    Launch date not confirmedDownloadable
  • A Muse Spark model that accepts text, images, video, audio and PDF through the API and Muse Code.

    Launch date not confirmedAPI · Provider product
  • A 30B dense model distilled from Muse Spark that accepts text and images and returns text.

    Launch date not confirmedDownloadable
  • A 16-expert Llama 4 model with 109B total and 17B active parameters.

    Launch date not confirmedDownloadable
  • A 128-expert Llama 4 model with 400B total and 17B active parameters.

    Launch date not confirmedDownloadable
  • An earlier Muse Spark API model with a 1,048,576-token context and Standard data mode.

    Launch date not confirmedAPI
  • The middle-sized Llama 3.1 instruction model, with 70B parameters.

    Launch date not confirmedDownloadable
  • The larger Llama 3.2 text assistant, with 3.21B parameters.

    Launch date not confirmedDownloadable
  • A compact Llama 3.2 text assistant with 1.23B parameters.

    Launch date not confirmedDownloadable
  • A 25-language transcription model for live and recorded audio with speaker labels and vocabulary hints.

    Launch date not confirmedAPI
  • A SAM update that uses Object Multiplex and shared memory to track several objects jointly.

    Launch date Downloadable · API
  • The largest Llama 3.1 text instruction model, with 405B parameters.

    Launch date not confirmedDownloadable
  • An 11B vision-language model that adds an image adapter to the Llama 3.1 text backbone.

    Launch date not confirmedDownloadable
  • A 90B image-and-text instruction model in the Llama 3.2 family.

    Launch date not confirmedDownloadable
  • Muse Image 1.0

    Meta · Muse Image

    An image generation and editing model that accepts text and reference images in a conversation.

    Launch date not confirmedAPI
  • A preview model for complex coding and long-running agents that accepts text, images, video, audio and PDF.

    Launch date not confirmedPreviewAPI · Provider product
  • A higher-capability Muse Wonder variant offered exclusively through Muse Code.

    Launch date not confirmedPreviewProvider product
  • SAM 3

    Meta · SAM

    A segmentation model that detects, outlines and tracks objects from short text concepts or visual prompts.

    Launch date not confirmedDownloadable
  • SAM 3D Body

    Meta · SAM 3D

    A model that estimates a full-body human mesh from one image, including body, hand and foot pose.

    Launch date not confirmedDownloadable
  • A model that reconstructs object shape, texture, pose and layout from masked regions of one image.

    Launch date not confirmedDownloadable