Models

Models

Routify supports 20+ models. All are accessible via the OpenAI-compatible /v1/chat/completions endpoint by passing the model name below.

Chinese models

ModelContextInput $/MOutput $/MBest for
deepseek-v3.264k$0.14$0.28Code, chat — cheapest GPT-4-class
deepseek-r164k$0.55$2.19Math & step-by-step reasoning
kimi-k2.5128k$0.60$2.50Long context, tool calling
glm-4.6128k$0.50$1.50Tool calling, agent workflows
qwen3-max256k$0.60$2.40Long-context understanding
qwen3-coder128k$0.30$1.50Code generation
doubao-1.6-pro64k$0.40$1.20Multimodal (images)
minimax-m2.5200k$0.30$1.20Speed (130 t/s)

Global models

ModelContextInput $/MOutput $/MBest for
claude-opus-4-7200k$15$75Premium reasoning, code
claude-sonnet-4-6200k$3$15Cursor / Claude Code default
gpt-4o128k$2.50$10Multimodal
o1200k$15$60Hard reasoning
gemini-3-pro2M$1.25$5Massive context

Video

ModelPricingBest for
kling-1.6$0.50 / sec of generated videoText-to-video

Smart routing

You can also pass routify-auto and we pick the lowest-cost healthy channel for your prompt's class:

client.chat.completions.create(
    model="routify-auto",
    messages=[{"role": "user", "content": "Hello"}],
)

Default class is chat. Override with extra_body:

extra_body={"routify": {"task": "code"}}

Tasks: chat | code | reasoning | longContext | vision | video.

Region preference

Set in dashboard or per-request:

extra_body={"routify": {"region": "domestic"}}  # CN-only channels
extra_body={"routify": {"region": "overseas"}}  # global channels

Pricing tiers

Your account group determines the multiplier on the official rate:

GroupMultiplierNotes
free0× (free quota only)$5 free credit, then locked until upgrade
domestic_pro1.10×Standard CN pricing
domestic_vip1.05×Volume tier
overseas_pro1.80×Standard global pricing
overseas_vip1.50×Volume tier
enterpriseCustomContact us

The published $/M numbers above are at our cost basis; your actual rate is cost × group_multiplier. Contact hello@routify.bytedance.city for enterprise / volume pricing.