How much could you save?

Enter your monthly token usage and the model you currently pay for. We'll show your real cost on each Chinese frontier model — and what stays in your pocket. Prices in USD per 1M tokens.

YOUR SETUP

Current model

Input tokens / month

M tokens

Output tokens / month

M tokens

Quick presets

Current cost on GPT-5.5

$550

per month

BEST SAVINGS

MiniMax M2.5

$9/mo · $0.4/M output

$541/mo

−98% · $6,492/yr

🇨🇳MiniMax M2.5OPEN

$9/mo−98%

🇨🇳DeepSeek-V4-FlashOPEN

$10.3/mo−98%

🇨🇳GLM-5.2OPEN

$19.06/mo−97%

🇨🇳Qwen3 MaxOPEN

$22.3/mo−96%

🇨🇳DeepSeek-V4-ProOPEN

$30.88/mo−94%

🇨🇳Doubao-Seed-2.0-Pro（豆包2.0 Pro）

$47.06/mo−91%

🇨🇳Kimi K2.5OPEN

$66/mo−88%

🇨🇳Kimi K3OPEN

$317.65/mo−42%

Estimates based on list prices (USD / 1M tokens) as of the data date. Actual cost varies with cache-hit rates, batch discounts and provider. Open-weight models can be self-hosted for compute cost only — often even cheaper at scale.

← Back to Model Index