Choose Auto, TokenLab Verified, or Official for each request, with prices shown up front. See what's new

ByteDance: Doubao-Seed-Evolving

Rolling ByteDance Seed model for rapidly updated reasoning, coding, and agent capabilities
Compare models
doubao-seed-evolving
AvailableByteDanceChat
Input / Output
$0.8824 / $4.41
Context
256K
Max output
256K
Modalities
VisionChat
Capabilities
Tool usePrompt CacheReasoning

About Doubao-Seed-Evolving

Doubao-Seed-Evolving is ByteDance's rolling deep-thinking model, announced at the Volcano Engine FORCE conference in June 2026 for agent and coding work. Its model ID stays fixed while ByteDance updates the model behind it, so calling it always reaches the latest version. It takes text and images, reasons, calls tools, and handles complex task orchestration and long-range planning.

Where it works well

  • A single ID tracks ByteDance's newest Seed capabilities without code changes, because the model behind it is updated on a rolling basis.
  • Aimed at complex task orchestration and long-range planning, the work that agent frameworks spend most of their effort on.
  • Handles code generation and tool invocation, the two skills most agent loops depend on.
  • Accepts images with text and offers a reasoning mode.
  • Tool calling and prompt caching are enabled.

When to choose another model

  • Because the model changes underneath the same ID, outputs can shift between weeks, which complicates regression tests and reproducibility.
  • Schema-enforced JSON output is not available; use a pinned Seed 2.1 model for strict schema output.
  • A pinned model such as Seed 2.1 Pro is easier to evaluate when you must prove stable behaviour.

Getting started

  1. Create API key

    Create a key in Console, then use it with every model on the platform.

  2. Send your first request

    Copy the example for your language and run it against the endpoint.

    Text to textResponses API
    POST/v1/responses
    API format:
    curl https://api.tokenlab.sh/v1/responses \
      -H "Content-Type: application/json" \
      -H "Authorization: Bearer sk-xxx" \
      -d '{
        "model": "doubao-seed-evolving",
        "input": "Hello!"
      }'

Pricing

The Verified price applies to Verified, which costs less on most models. The Official price is the model maker's published price and applies to the more reliable Official route. Auto bills the route that completes the request.

Input tokens <= 256K · Text to text

per 1M tokens
Official price
Input $0.8824 / Output $4.41 / Cache read $0.1765
Verified price
—
Discount
—
Prompt cache pricing

Cache read

Official price
$0.1765
Verified price
—
Discount
—

Usage & activity

Success rate is the share of requests that completed. Latency is how long a full response takes; P95 means 95% of requests finished within that time.

Usage & availability

Last 24 hours
Requests
Success rate
P95 latency
Total tokens
Model performance
Metrics appear once privacy and data volume thresholds are met.

Data is based on aggregate user requests, excluding status checks.

Open in Console

Open Doubao-Seed-Evolving in Console with a prompt ready to edit or send.

Help me try doubao-seed-evolving with a short message at /v1/responses. Show the reply, latency, and cost.

Use cases

Best for
  • Reasoning
  • Vision
  • Exploring new capabilities

    Point an evaluation harness at the rolling ID to see how ByteDance's latest Seed improvements affect your own tasks.

  • Long-horizon planning agents

    Let it decompose a large goal into steps, call tools for each, and revise the plan as results arrive.

  • Coding assistants that follow the frontier

    Keep a developer tool on current Seed coding ability without re-pinning each release.

  • Canary traffic

    Send a small share of traffic to it and compare against a pinned model before adopting changes.

Prompt examples

Break this goal into a dependency-ordered plan: migrate our billing service to a new database with no downtime. Then start step one using the tools.

Review these three modules, find the one most likely to break under concurrent writes, and write a failing test for it.

The attached diagram shows our pipeline. Identify the bottleneck and propose two changes, ranked by effort.

This model has conditional pricing. Monthly token totals alone cannot produce a reliable estimate; use the detailed pricing for the request specification, cache, and applicable time window.

FAQ

What exactly is Doubao-Seed-Evolving?

It is a rolling model from ByteDance's Seed line, introduced in June 2026 for agent and coding scenarios. The ID stays constant while the underlying model is updated, so you always call the latest version.

How often does it change?

ByteDance updates the model behind the ID on a rolling basis, often several times a month. Treat results as time-dependent and re-run your tests regularly.

Should I use it in production?

Use it where gaining the newest capability matters more than exact repeatability. If you need stable, auditable behaviour, a pinned model like Doubao-Seed-2.1-Pro is the safer base.

Does it support images and tools?

Yes. It accepts images with text, reasons before answering, and supports tool calling and prompt caching. Schema-bound JSON output is not supported, so validate structured replies yourself.

How does it differ from Seed 2.1 Pro?

Seed 2.1 Pro is a fixed release with published benchmarks. Evolving is a moving target that tracks newer improvements, so its behaviour has no fixed reference point.

How much does Doubao-Seed-Evolving cost?

On TokenLab, Doubao-Seed-Evolving costs Input $0.8824 / Output $4.41 per 1M tokens. The pricing table above shows the full breakdown. Rates depend on the billing unit, specification, and usage. Compare matching conditions in the model's detailed pricing; a single rate does not determine the total cost.

What are the context window and output limit of Doubao-Seed-Evolving?

Doubao-Seed-Evolving accepts up to 262,144 tokens of context and returns up to 262,144 tokens in one response.

Which endpoint should Doubao-Seed-Evolving use?

Use https://api.tokenlab.sh/v1/responses for Doubao-Seed-Evolving. The request example below shows the matching code shape.

Which operations does Doubao-Seed-Evolving support?

Doubao-Seed-Evolving supports Text to text. Select an operation above to see its endpoint and request example.

Compare Doubao-Seed-Evolving

Sources

Reviewed Oct 2, 2026

More from Doubao

Related models