<!-- Generated from public website sources; do not edit this file. -->

> DSSUB design reference. Account, payments and API are not connected; prices, policies and setup examples require independent launch verification.

# Plan

This is a design preview. Account, payment and API services are not connected. You can browse model details, compare design prices, calculate workloads, download configuration templates and explore sample account data. No real key, order or API request is created.

[View service availability](/service-status/index.md).

## DeepSeek Plan

| Plan | One-time month (USD) | Monthly charge (USD) | Annual charge (USD) | 5-hour allowance | 7-day allowance |
| --- | --- | --- | --- | --- | --- |
| Base | $29.90 | $19.90 | $179.00 | 60M | 400M |
| Max | $59.90 | $39.90 | $359.00 | 150M | 1,000M |
| Pro Max | $119.90 | $79.90 | $719.00 | 400M | 2,800M |
| Ultra | $499.00 | $329.00 | $2990.00 | No 5-hour cap | 8,400M |

Annual amounts are charged upfront, not in monthly instalments. M means one million plan points.

## Prefer to pay by token?

A subscription includes a plan allowance for a monthly or yearly fee. With prepaid API credit, you pay for input, cached input and output tokens at the listed model rates. API credit is separate from your subscription allowance.

[View API pricing](/points/#usage-prices-title)

## DeepSeek Plan · Prices and allowances

Design prices and usage windows are listed below. Purchases are unavailable. Final products, charges and delivery terms must be confirmed before checkout opens.

| Tier | One-time · 1 month | Monthly | Annual total | 5h / 7d allowance |
| --- | --- | --- | --- | --- |
| Base | $29.90 once | $19.90 / month | $179.00 / year | 60M / 400M |
| Max | $59.90 once | $39.90 / month | $359.00 / year | 150M / 1,000M |
| Pro Max | $119.90 once | $79.90 / month | $719.00 / year | 400M / 2,800M |
| Ultra | $499.00 once | $329.00 / month | $2990.00 / year | No 5-hour cap / 8,400M |

M means one million plan points. One-time access lasts one calendar month without renewal. Annual billing is paid upfront. Discounts are relative to one-time purchases for the same duration. One-time, monthly and annual purchases include the same allowances. Use deepseek-flash or deepseek-v4-pro. Flash uses 1 plan token per token; Pro uses 3.5. Rate and concurrency limits apply. Your account shows allowance usage and reset times. Ultra has no 5-hour quota cap and uses the 7-day allowance shown above.

### Model pricing & points

DeepSeek V4 Pro: 3.5× the price of DeepSeek V4 Flash; 3.5× the points consumption. Flash uses 1 point per token; V4 Pro uses 3.5 points per token.

For example, 1M V4 Pro tokens consume 3.5M points. The allowances shown use Flash as the baseline.

Proposed terms: subscriptions would renew until cancelled. The design includes refund review within 15 calendar days; final terms and exceptions must be confirmed before launch. [Read cancellation and refund terms](/refunds/index.md) and [usage limits](/docs/usage-limits/index.md).

## Multi-Model Plan · Coming soon

Planned access to DeepSeek, GLM, Kimi, MiMo, Qwen and MiniMax. Purchases and model access are not available yet. Multi-Model has its own allowances; a DeepSeek Plan does not include Multi-Model access.

| Tier | One-time · 1 month | Monthly | Annual total | 5h / 7d allowance |
| --- | --- | --- | --- | --- |
| Base | $29.90 once | $19.90 / month | $179.00 / year | 60M / 300M |
| Max | $59.90 once | $39.90 / month | $359.00 / year | 150M / 900M |
| Pro Max | $119.90 once | $79.90 / month | $719.00 / year | 400M / 2,400M |
| Ultra | $499.00 once | $329.00 / month | $2990.00 / year | No 5-hour cap / 5,000M |

M means one million plan points. Per-model point rates will be published at launch.

## Usage and limits

Understand the design allowances and the distinction between tokens and plan points.

### Current scope
These are design rules for the DeepSeek Plan preview. No production gateway, quota enforcement or billing service is connected. Final access and reset behavior must be verified before launch.

### Plan allowances
| Plan | 5-hour allowance | 7-day allowance |
| --- | --- | --- |
| Base | 60M points | 400M points |
| Max | 150M points | 1,000M points |
| Pro Max | 400M points | 2,800M points |
| Ultra | No 5-hour cap | 8,400M points |

M means one million plan points. One-time, monthly and yearly designs share these allowances. Rate and concurrency limits remain separate and are not yet confirmed for production.

### Count tokens once
Flash uses 1 point per input or output token; Pro uses 3.5. Cached input is part of input tokens, not an additional token total. For example, 1M Pro tokens consume 3.5M points in the design model.

### Check both windows
The design uses fixed 5-hour and 7-day windows anchored to activation. A burst can exceed a short window even when the month looks affordable. The [calculator](/compare/#cost-calculator) checks your stated peaks and a normalized 30-day allowance; actual timing and window boundaries still matter.

### Read usage and settlement
The [Usage preview](/account/usage) separates raw tokens, plan points and USD API charges. Its charts cover the applied period; request history has filters, pagination, details and current-page CSV export. Cached input is already included in input totals. A reserved request is not a finalized charge.

### Billing and renewal
The design includes one calendar month without renewal, recurring monthly billing and annual billing paid upfront. Prices remain on the [Plan page](/subscription/index.md). No payment, renewal or cancellation is active. The [Billing preview](/account/billing) illustrates payment, delivery and renewal states separately.

See [cancellation and refunds](/refunds/index.md), [service availability](/service-status/index.md) and [support](/support/index.md) for current scope.

---

HTML source: [Plan — DeepSeek & Multi-Model](/subscription/)
