Here are the cheapest capable Chinese LLM APIs in 2026, ranked by input price. All prices are the upstream rate — NovAI adds 0% platform markup.
| # | Model | Input | Output | Provider |
|---|---|---|---|---|
| 1 | qwen3.5-flash | $0.063 | $0.336 | Alibaba |
| 2 | doubao-seed-2.0-lite | $0.075 | $0.45 | ByteDance |
| 3 | hy3-preview | $0.127 | $0.423 | Tencent |
| 4 | doubao-seed-2.0-code | $0.448 | $2.24 | ByteDance |
| 5 | deepseek-v4-flash | $0.168 | $0.336 | DeepSeek |
| 6 | hy3 | $0.18 | $0.72 | Tencent |
| 7 | qwen3.5-plus | $0.234 | $1.404 | Alibaba |
| 8 | minimax-m2.5 | $0.282 | $1.128 | MiniMax |
And glm-4.6v-flash is completely free — $0 input and output.
from openai import OpenAI
client = OpenAI(
api_key="nvai-your-key",
base_url="https://aiapi-pro.com/v1"
)
resp = client.chat.completions.create(
model="qwen3.5-flash",
messages=[{"role": "user", "content": "Hello!"}]
)
print(resp.choices[0].message.content)All the cheapest Chinese models on one endpoint — 0% markup, $2 free credit.
Qwen3.5 Flash at $0.063 per 1M input tokens is the cheapest capable option, followed by Doubao-Seed-2.0-Lite ($0.075) and Hunyuan hy3-preview ($0.127). GLM-4.6V Flash is completely free.
Yes. GLM-4.6V Flash is free on NovAI with no token limit and no credit card. It's ideal for prototyping before moving to a paid model.
No. NovAI charges 0% platform markup — the prices shown are the upstream provider rates. New accounts get $2 free credit that never expires.