Back to Models

Doubao Seed 2.0 Lite

doubaodoubao/doubao-seed-2-0-lite

Doubao Seed 2.0 Lite β€” cost-efficient Seed 2.0 variant.

Doubao-Seed-2.0-lite (Doubao Seed 2.0 Lite) is ByteDance's cost-efficient mid-tier model in the Seed 2.0 family, released February 14, 2026 via Volcano Engine. It is positioned between the flagship Pro variant and the ultra-light Mini, targeting balanced performance and throughput for production Agent workloads. The model ships with 131 k context and 32 k max completion and is fully OpenAI-compatible through api.therouter.ai/v1.

On May 6, 2026 the model received a significant upgrade that made it the first omni-modal understanding model in the Doubao family, adding native video, image, and audio understanding alongside text. Agent, coding, and GUI capabilities were also improved.

Best for
  • β€’ Cost-sensitive production Agent workloads that need stronger reasoning than Mini but lower latency/cost than Pro (customer service Q&A, information extraction, K-12 tutoring, intent recognition).
  • β€’ High-QPS edge deployments where 131 k context and 32 k generation suffice and the operator wants predictable pricing with cache-read discounts.
Reach for something else if
  • β€’ Frontier research-level reasoning or long-horizon scientific discovery β€” route to doubao/doubao-seed-2-0-pro or anthropic/claude-opus-4.7 instead.
  • β€’ Native video or audio understanding workloads β€” currently only available directly via Volcano Engine, not yet routed through TheRouter.

How TheRouter serves this differently from the vendor

As the vendor operates it

Omni-modal understanding: native video, image, and audio input alongside text, available directly through Volcano Engine.

On TheRouter

Text in, text out. The omni-modal inputs are not routed here yet, so send text only.

Context Length
131K
Max Output
33K
Input Priceper 1M tokens
$0.270/ 1M tokens
Output Priceper 1M tokens
$2.16/ 1M tokens

Modalities

text→text

Pricing Breakdown

TypeRate
Input$0.270 / 1M tokens
Output$2.16 / 1M tokens

Supported Parameters

temperaturemax_tokenstop_ptoolstool_choiceresponse_formatstop

Specifications

Release date2026-02-14seed.bytedance.com β†—verified
Omni-modal upgrade date2026-05-06www.yicai.com β†—verified
LicenseProprietary (Volcano Engine API)verified
Training cutoffNot publicly disclosedunknown

API Usage Examples

Use the global api.therouter.ai endpoint shown below for new integrations; the legacy China accelerated endpoint is retired.

cURL
curl https://api.therouter.ai/v1/chat/completions   -H "Content-Type: application/json"   -H "Authorization: Bearer $THE_ROUTER_API_KEY"   -d '{
    "model": "doubao/doubao-seed-2-0-lite",
    "messages": [
      {"role": "user", "content": "Summarize the key points from this input."}
    ]
  }'

API guide

Chat completion

cURL
curl https://api.therouter.ai/v1/chat/completions \
  -H "Authorization: Bearer $THEROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"doubao/doubao-seed-2-0-lite","messages":[{"role":"user","content":"Hello"}]}'

More from doubao

Similar models

Cross-provider sibling models

News & changes

2026-05-06

Doubao-Seed-2.0-lite upgraded to first omni-modal model in Doubao family

Volcano Engine announced native unified understanding of video, image, audio and text, plus Agent/Coding/GUI capability improvements. The upgrade positions Lite as the most cost-effective omni-modal option in the Seed 2.0 lineup.

re-authored by TheRouteryicai.com β†—
2026-02-14

ByteDance releases Seed 2.0 series β€” Pro positioned against GPT-5.2 / Gemini 3 Pro

Seed 2.0 Pro, Lite, Mini and Code variants launched simultaneously on Volcano Engine API. Pro claims frontier-level results on multimodal, math and coding benchmarks at ~10Γ— lower token price than Western frontier models.

re-authored by TheRouterseed.bytedance.com β†—

Frequently asked

Does Doubao-Seed-2.0-lite support native video or audio understanding through TheRouter?

No. The May 2026 omni-modal upgrade is currently only available via direct Volcano Engine endpoints. TheRouter's standard-models.yaml still lists the model as text-only. Operators who need video/audio understanding should call Volcano Engine directly or wait for TheRouter to expose the expanded modality.

re-authored by TheRouterwww.yicai.com β†—
How does Seed 2.0 Lite compare to Seed 2.0 Mini and Pro in cost/performance?

Lite sits between Mini (lowest cost, lowest latency, high-QPS edge) and Pro (strongest reasoning, highest cost). It offers the best balance for most production Agent workloads that need more capability than Mini but do not justify Pro pricing.

re-authored by TheRouterseed.bytedance.com β†—
Fact ledger β€” every claim on this page traces here
sourceURLretrieved
Release dateseed.bytedance.com β†—2026-05-25verified
Omni-modal upgrade datewww.yicai.com β†—2026-05-25verified
Licenseβ€”β€”verified
Training cutoffβ€”β€”unknown
Doubao-Seed-2.0-lite upgraded to first omni-modal model in Doubao familyyicai.com β†—2026-05-25verified
ByteDance releases Seed 2.0 series β€” Pro positioned against GPT-5.2 / Gemini 3 Proseed.bytedance.com β†—2026-05-25verified
Does Doubao-Seed-2.0-lite support native video or audio understanding through TheRouter?www.yicai.com β†—2026-05-25to verify
How does Seed 2.0 Lite compare to Seed 2.0 Mini and Pro in cost/performance?seed.bytedance.com β†—2026-05-25to verify
Help & contact