Pay for the tokens you use, or choose a Coding Plan with shared points. All prices below are in USD.
FIVE PROVIDERS. ONE PLAN.
Switch models. Keep one allowance.
Choose the allowance that fits your usage. All three plans include DeepSeek, GLM, Kimi, MiniMax, and Qwen with the same API key.
Save about 50% with automatic renewalPrices in USD
Pro
For regular coding across your preferred models.
$29.90/ month
Billed every month. Cancel anytime.
$59.90 / month · Same allowance and points rates
Shared points / 7 days ?
Based on GLM 5.3 Flash: 1 input or output token uses 1 point. GLM 5.3 uses 9 points per token; other models use their listed rates. Usage counts against both fixed windows, starting from plan activation. Monthly subscriptions reduce the plan fee by about 50%; allowances and deduction rates are unchanged. See model rates and reset rules →
300M
Shared points / 5 hours ?
Based on GLM 5.3 Flash: 1 input or output token uses 1 point. GLM 5.3 uses 9 points per token; other models use their listed rates. Usage counts against both fixed windows, starting from plan activation. Monthly subscriptions reduce the plan fee by about 50%; allowances and deduction rates are unchanged. See model rates and reset rules →
60M
All 5 model providers
Switch models with the same key
OpenAI & Anthropic-compatible APIs
Tool-specific setup guides
Max
Recommended
More weekly allowance for longer coding sessions.
$49.90/ month
Billed every month. Cancel anytime.
$99.90 / month · Same allowance and points rates
Shared points / 7 days ?
Based on GLM 5.3 Flash: 1 input or output token uses 1 point. GLM 5.3 uses 9 points per token; other models use their listed rates. Usage counts against both fixed windows, starting from plan activation. Monthly subscriptions reduce the plan fee by about 50%; allowances and deduction rates are unchanged. See model rates and reset rules →
800M
Shared points / 5 hours ?
Based on GLM 5.3 Flash: 1 input or output token uses 1 point. GLM 5.3 uses 9 points per token; other models use their listed rates. Usage counts against both fixed windows, starting from plan activation. Monthly subscriptions reduce the plan fee by about 50%; allowances and deduction rates are unchanged. See model rates and reset rules →
120M
All 5 model providers
Switch models with the same key
OpenAI & Anthropic-compatible APIs
Tool-specific setup guides
Pro Max
The largest allowance for sustained, high-volume work.
$149/ month
Billed every month. Cancel anytime.
$299 / month · Same allowance and points rates
Shared points / 7 days ?
Based on GLM 5.3 Flash: 1 input or output token uses 1 point. GLM 5.3 uses 9 points per token; other models use their listed rates. Usage counts against both fixed windows, starting from plan activation. Monthly subscriptions reduce the plan fee by about 50%; allowances and deduction rates are unchanged. See model rates and reset rules →
2,800M
Shared points / 5 hours ?
Based on GLM 5.3 Flash: 1 input or output token uses 1 point. GLM 5.3 uses 9 points per token; other models use their listed rates. Usage counts against both fixed windows, starting from plan activation. Monthly subscriptions reduce the plan fee by about 50%; allowances and deduction rates are unchanged. See model rates and reset rules →
Enter tokens for one request or a whole workload. Values are in millions; cached input is separate from uncached input.
Estimated API usage cost$0.685Same workload on Coding Plan2M points
Input $0.135 + cached input $0.00 + output $0.55.
Estimate uses the listed token rates, before tax or credit rewards. Actual cache hits come from reported usage.
Questions about your plan
What do I get with SI-API?
One API key for DeepSeek, GLM, Kimi, MiniMax, and Qwen. Use prepaid API credit at each model’s token prices, or choose an optional Coding Plan with shared points.
Can I switch models on the same plan?
Yes. All seven listed models share your plan allowance. Change the model ID in your tool and keep the same key. Each model uses points at its listed rate.
Should I choose a Coding Plan or pay as you go?
Choose a Coding Plan for regular coding when its 5-hour and 7-day allowances fit your workload. Choose pay as you go for occasional use or variable workloads billed by input, cached input, and output tokens.
How do the two plan allowances work?
Every request counts toward both the 5-hour and 7-day allowances. Both need available capacity. Quota resets are separate from your monthly subscription or one-time access period.
What happens when my plan runs out?
Requests pause by default. You can enable pay-as-you-go fallback in Billing to continue with available API credit. You control whether your balance is used.
How do I connect my coding tool?
Create your API key and choose your Agent from seven tool guides. Each setup includes all seven SI-API models. Configure it once, then switch models inside your Agent using the same key and endpoint. For applications, use the API quickstart.
Can I buy a month without renewing?
Yes. Choose a one-time purchase for one calendar month. A monthly subscription renews automatically at its displayed price and can be canceled in Billing.
