Skip to content

豆包大模型

A ByteDance large model family with multimodal capabilities

Visit official site

What 豆包大模型 is

The model family is described as a self-developed large model from ByteDance with multimodal capabilities, validated through more than 50 internal business scenarios and daily usage at the trillion-token scale. The material states the family covers general-purpose, video generation, text-to-image, image-to-image and simultaneous interpretation models for different business needs.

What you can do with it

  • Use the Pro general model for QA, summaries, creation and classification
  • Use the Lite general model where lower token cost and latency matter
  • Create videos from text or images with the video generation model
  • Generate images from text or transform photos across many styles

Model notes

  • The Pro general model supports 128K long text and full-series fine-tuning
  • The Lite general model is aimed at scenarios with high real-time requirements
  • The image-to-image model supports more than 50 style transformations
  • The simultaneous interpretation model supports cross-language translation with the same voice timbre

Application scenarios

  • Enterprises use the Pro general model for content creation and knowledge QA
  • Media companies use the video generation model for news report videos
  • Education providers use the simultaneous interpretation model for live course translation

Pros & cons

✓ What we like

  • The family covers general, video, image and interpretation models
  • Multimodal capability is described across the model family

! What to watch out for

  • The documented setup includes reading usage limits and billing methods
  • Model choice follows the business need described in the guide

FAQ

Which models are in the family?

The material lists general-purpose Pro and Lite, video generation, text-to-image, image-to-image and simultaneous interpretation models.

How long is the Pro context?

The feature list states the Pro general model supports 128K long text and full-series fine-tuning.

Last reviewed: 2026-11-15

More LLM API platform tools

View all →

How we review