# Is the DeepSeek API free? No, it bills per token from $0.15 per 1M

> DeepSeek's docs describe no free tier or credit amount. Usage is deducted from a topped-up or granted balance, and requests fail with a 402 when it hits zero.

URL: https://claudegpt.si/news/is-deepseek-api-free  
Published: 2026-11-02T06:55:09.377Z  
Author: ClaudeGPT Si Newsroom  
Section: AI Models  
Topics: DeepSeek, API pricing, AI pricing, Language models  
Publisher: ClaudeGPT Si

## Key takeaways

- DeepSeek's API docs describe no free tier. Requests are billed per token from a topped-up or granted balance.
- The docs do not say who gets a granted balance or how large it is. Running out returns HTTP 402, Insufficient Balance.
- As of November 2, 2026, deepseek-flash costs $0.15 per 1M input tokens (cache miss) and $0.60 per 1M output tokens off-peak.

Is the DeepSeek API free? No. DeepSeek's [pricing page](https://api-docs.deepseek.com/quick_start/pricing) bills every request by the token, and it does not promise free credits or a free tier. As of November 2, 2026, the cheapest model, `deepseek-flash`, costs $0.15 per 1 million input tokens off-peak when the prompt is not cached.

That is cheap, and it is easy to confuse with free. This piece covers what the docs say about balances, what the API costs, and what happens when you run out. For the chat app, see our answer on [whether DeepSeek is free](/news/is-deepseek-free).

## Does DeepSeek give free API credits?

DeepSeek's docs never say. The pricing page says fees "will be directly deducted from your topped-up balance or granted balance," and that the granted balance is used first when you have both. It does not say what a granted balance is, who receives one, how large it is or when it expires.

We also checked the API quick start, the rate limit page and the error codes page. None of them mention free tokens, a free tier or a sign-up bonus. So the safe reading is that the API is a paid product, and that some accounts may hold a granted balance whose terms DeepSeek has not published in these docs. Check your own balance in your account before you count on any.

You do need an API key to call the API at all, per the [quick start](https://api-docs.deepseek.com/).

## What happens when your balance runs out

The [error codes page](https://api-docs.deepseek.com/quick_start/error_codes) lists HTTP 402, "Insufficient Balance." DeepSeek's stated cause is "You have run out of balance," and the fix is to add funds on the Top up page. There is no free fallback.

## DeepSeek API pricing

Prices are per 1 million tokens in US dollars, from the pricing page as of November 2, 2026. DeepSeek reserves the right to change them.

| Per 1M tokens | Flash, off-peak | Flash, peak | V4 Pro, off-peak | V4 Pro, peak |
| --- | --- | --- | --- | --- |
| Input, cache hit | $0.003 | $0.006 | $0.022 | $0.044 |
| Input, cache miss | $0.15 | $0.30 | $0.66 | $1.32 |
| Output | $0.60 | $1.20 | $1.98 | $3.96 |

A cache hit means the start of your prompt matches one DeepSeek processed recently, so reading it again costs less. A cache miss is a prompt it reads fresh.

Both models have a 1M-token context window and can write up to 384K tokens per reply. `deepseek-flash` runs DeepSeek-V4.1-Flash-1017 and accepts images. `deepseek-v4-pro` runs DeepSeek-V4-Pro-0813 and does not.

### A rough sense of cost

Expense is tokens times price. At the Flash off-peak cache-miss rate, a prompt of 1 million tokens costs $0.15 to read, and 1 million tokens of output costs $0.60. Small tests use a tiny fraction of that, but they are still billed.

## Peak and off-peak hours

Off-peak rates are half the peak rates. Peak hours are 01:00 to 04:00 and 06:00 to 10:00 UTC, Monday to Friday, excluding Chinese public holidays. Every other hour, including weekends, is off-peak.

## Search and fetch are billed per request

DeepSeek's standalone Search and Fetch APIs each cost $0.75 per 1,000 successful requests. Server-side tool calls made through DeepSeek Harness, Claude Code or Codex are currently free of charge. Per the pricing page, per-request billing starts at 00:00 Beijing time on November 1, 2026, so check the page for the current status before you rely on that.

## Free limits do not apply, but rate limits do

The [rate limit page](https://api-docs.deepseek.com/quick_start/rate_limit) caps concurrent requests per account at 2,500 for `deepseek-flash` and 500 for `deepseek-v4-pro`. Going over returns HTTP 429. DeepSeek says higher concurrency is available on request at no additional cost. These are limits on speed, not a free allowance.

## What to do next

If you want to test without paying much, top up a small amount, call `deepseek-flash` off-peak and watch the balance. Confirm the price on the [pricing page](https://api-docs.deepseek.com/quick_start/pricing) first, because it can change.

## Frequently asked questions

### Does DeepSeek API have a free tier?

Not in its docs. The pricing, quick start, rate limit and error code pages describe paid, per-token billing and mention no free tier or free tokens.

### Does DeepSeek give free API credits to new users?

DeepSeek does not say. The pricing page mentions a granted balance that is used before a topped-up balance, but does not explain who gets it or how much it is.

### What happens when my DeepSeek API balance runs out?

Requests return HTTP 402, Insufficient Balance. DeepSeek's fix is to add funds on the Top up page.

### How much does the DeepSeek API cost?

As of November 2, 2026, deepseek-flash costs $0.15 input (cache miss) and $0.60 output per 1M tokens off-peak. deepseek-v4-pro costs $0.66 and $1.98. Peak rates are double.

## Sources

- [Models & Pricing](https://api-docs.deepseek.com/quick_start/pricing), DeepSeek API Docs
- [Error Codes](https://api-docs.deepseek.com/quick_start/error_codes), DeepSeek API Docs
- [Rate Limit & Isolation](https://api-docs.deepseek.com/quick_start/rate_limit), DeepSeek API Docs
- [Your First API Call](https://api-docs.deepseek.com/), DeepSeek API Docs