Back to Models

Doubao 1.5 Lite 32k

doubaodoubao/doubao-1-5-lite-32k

Doubao 1.5 Lite (32k context) β€” cost-efficient ByteDance chat model for high-volume routine tasks.

Doubao-1.5-lite-32k is the lightweight, cost-optimized member of ByteDance's Doubao 1.5 model family, launched January 22, 2025 on Volcano Engine. It targets high-volume, latency-sensitive workloads where strong responsiveness and low unit cost matter more than frontier reasoning depth. The 1.5 Lite inherits the sparse MoE architecture and independent data-production pipeline used across the 1.5 series, delivering competitive lightweight performance at roughly one-third the price of Doubao-1.5-pro-32k.

For TheRouter operators, Doubao-1.5-lite-32k is the default low-cost Chinese-language chat endpoint when traffic volume is high and per-token spend must stay minimal. It supports the same OpenAI-compatible parameter surface (temperature, top_p, tools, response_format, stop) as the rest of the Doubao family, making it a drop-in routing target for cost-tiering strategies alongside the heavier 1.5-pro and Seed-series models.

Best for
  • β€’ High-volume, latency-sensitive Chinese chat and classification workloads where cost per million tokens must stay below ~$0.10.
  • β€’ Cost-tier fallback or shadow traffic alongside Doubao-1.5-pro-32k and Seed 1.6 when per-request budget is constrained.
Reach for something else if
  • β€’ Complex multi-step reasoning, long-context document analysis, or English code generation β€” use Doubao-1.5-pro-32k, Seed 1.6, or Qwen3-235B instead.
Context Length
33K
Max Output
16K
Input Priceper 1M tokens
$0.0486/ 1M tokens
Output Priceper 1M tokens
$0.0972/ 1M tokens

Modalities

text→text

Pricing Breakdown

TypeRate
Input$0.0486 / 1M tokens
Output$0.0972 / 1M tokens

Supported Parameters

temperaturemax_tokenstop_ptoolstool_choiceresponse_formatstop

Specifications

Release date2025-01-22 (Doubao 1.5 family)www.aibase.com β†—verified
ArchitectureSparse MoE (same family as 1.5-pro)seed.bytedance.com β†—verified
LicenseProprietary (API only)verified
Training cutoffNot publicly disclosedunknown
Data independenceNo synthetic data from other models used in trainingwww.aibase.com β†—verified

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-1-5-lite-32k",
    "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-1-5-lite-32k","messages":[{"role":"user","content":"δ½ ε₯½οΌŒθ―·η”¨δΈ­ζ–‡δ»‹η»θ‡ͺ己。"}]}'

More from doubao

Similar models

Cross-provider sibling models

News & changes

2025-01-22

ByteDance releases Doubao Model 1.5 family β€” "more features, same price"

Volcano Engine launched the full Doubao 1.5 series (Pro, Lite, Vision-Pro) on January 22, 2025. The Lite variant targets high-volume, cost-sensitive workloads while maintaining performance parity with the prior Pro-32k generation. All models share the same sparse-MoE backbone and independent data-production pipeline that avoids synthetic data from other LLMs.

re-authored by TheRouteraibase.com β†—
Fact ledger β€” every claim on this page traces here
sourceURLretrieved
Release datewww.aibase.com β†—2026-05-25verified
Architectureseed.bytedance.com β†—2026-05-25verified
Licenseβ€”β€”verified
Training cutoffβ€”β€”unknown
Data independencewww.aibase.com β†—2026-05-25verified
ByteDance releases Doubao Model 1.5 family β€” "more features, same price"aibase.com β†—2026-05-25verified
Help & contact