Skip to main content
Doubao-Seed-Evolving gives you ByteDance’s strongest available Seed model through a single, stable model ID, available through AnyFast via an OpenAI-compatible interface. It updates on a weekly cadence and continuously evolves — upgrades take effect automatically, so you never need to switch model IDs or rework your integration. It currently maps to Doubao-Seed-2.1-Pro, and targets coding, office, and productivity scenarios.
Doubao-Seed-Evolving does not support batch inference. For batch workloads, use a versioned model such as doubao-seed-2.1-pro or doubao-seed-2.1-turbo.

Key capabilities

  • OpenAI-compatible — Works as a drop-in replacement with the OpenAI SDK
  • Always latest — One model ID points to the strongest version, updated weekly with no migration
  • 256K context window — Handles project-scale context, with up to 256K tokens of output in a single response
  • Deep thinking — Chain-of-thought reasoning, on by default, with adjustable reasoning_effort
  • Multimodal understanding — Text, image, video, and document understanding (audio not supported)
  • Function calling, structured output & context caching — Robust tool use, JSON output (beta), and prompt caching
  • Streaming — Real-time token streaming via SSE

Output specifications

Quick example

Thinking mode

Doubao-Seed-Evolving reasons before answering by default. thinking.type is enabled unless you set it to disabled to skip reasoning for lightweight tasks. Use reasoning_effort to tune how long the model thinks.
Python
reasoning_effort controls reasoning length (effective only when thinking is enabled) and accepts minimal, low, medium, and high. Default: high.

Parameters

API Reference

View the interactive API playground for Doubao Seed Evolving.