Supported Models

All models are OpenAI API compatible. Prices in USD per 1M tokens, inclusive of 40% healthy margin.

Overseas FlagshipPrimary

ModelContextInput / 1M tokensOutput / 1M tokens
gpt-4o
Most capable multimodal
128K$15.55$62.22
gpt-4o-mini
Fast and affordable
128K$1.95$7.78
gpt-4-turbo
High-performance reasoning
128K$13.61$54.44
claude-3-opus
Flagship reasoning
200K$31.11$155.55
claude-3-sonnet
Best coding performance
200K$15.55$77.78
claude-3-haiku
Fast and cost-effective
200K$3.89$19.44
gemini-3.1-pro
Ultra-long context
2M$13.61$54.44
gemini-2.0-flash
Fast multimodal
1M$1.55$6.22
grok-4.20-beta
X.ai model
128K$17.50$70.00

Domestic ModelsNative

ModelContextInput / 1M tokensOutput / 1M tokens
qwen-3-max
Alibaba flagship
128K$5.84$23.34
deepseek-r1
Reasoning model
64K$4.86$19.44
glm-5
Zhipu AI
128K$5.44$21.78
doubao
ByteDance
128K$4.47$17.89
wenxin-4.0
Baidu ERNIE
128K$7.78$31.11
hunyuan
Tencent
128K$6.80$27.22

Low Cost Open SourceLow Cost

ModelContextInput / 1M tokensOutput / 1M tokens
llama-4-70b
Open-source via Groq
128K$1.20$3.50
mistral-large
Open-source via Groq
128K$1.40$4.20

Pricing Notes:

  • All prices in USD per 1 million tokens
  • Input and output tokens priced separately
  • Overseas models route with primary provider and auto-failover backup
  • Domestic models route via native line exclusively
  • Open-source models are user-selectable low-cost alternatives
  • Prices include 40% healthy margin, competitive vs official retail