What Claude Code costs per month on DeepSeek, GLM or Kimi
Updated 2026-10-18 · A typical month of agent steps priced model by model, and a calculator for your own numbers. Cached input is most of the bill.
Connection details
- Base URL (OpenAI compatible)
- https://api.aifans.fan/v1
- Base URL (Anthropic compatible)
- https://api.aifans.fan
- API key
- Create one in the console; it starts with sk- (no account yet? sign up)
- Recommended models
- deepseek-v4.1-flash, deepseek-v4-pro (all models and prices)
Claude Code sends the whole conversation with every step, so a coding session is mostly input tokens, and most of those are cache hits. That is why the price of cached input matters as much as the headline price. Below is a typical month and a calculator for your own numbers.
A typical month
200 agent steps a day for 22 working days, about 150K tokens of context per step of which 90% are cache hits, 1,000 output tokens per step, half of the calls in peak hours:
| Model | At the maker's own price | Here |
|---|---|---|
| DeepSeek V4.1 Flash | $21.48 | $12.83 |
| Hunyuan Hy3 | $30.64 | $25.98 |
| GLM-5.3 Flash | $29.92 | $28.28 |
| MiniMax M2.7 | $60.72 | $60.36 |
| DeepSeek V4 Pro | $98.01 | $63.17 |
| Hunyuan Hy4 Preview | $91.00 | $73.14 |
| GLM-5.3 | $266.20 | $251.62 |
| Kimi K3 | $442.20 | $439.92 |
Your numbers
How to spend less
- Run /compact now and then: a shorter context is cheaper on every step.
- Give the small, frequent jobs to a fast model (ANTHROPIC_DEFAULT_HAIKU_MODEL) and keep the strong model for the main work.
- Peak hours only matter for models with time-of-day prices; see peak hours in your time zone.
- Online top-ups carry a payment processing fee of 5.5% (minimum $0.8), added at checkout and shown before you pay; it is not part of your credit.
Estimates from list prices; your bill counts the actual tokens of each call, itemised in the console.