DeepSeek

What it does

DeepSeek is a Chinese AI company whose assistant answers questions, writes, reasons through problems and helps with code. Consumers use it in a browser at chat.deepseek.com or in the iOS and Android apps; the web chat offers an "Expert" and an "Instant" mode. Developers can call the same model family through the DeepSeek API, which provides two models, deepseek-flash and deepseek-v4-pro, each with a context window of one million tokens and both a thinking and a non-thinking mode; only the Flash model accepts images. DeepSeek published the weights of its V4 Preview models on Hugging Face; its release note does not name a licence.

Pricing and plans

DeepSeek's download page describes the chat app as free. There is no consumer subscription listed. The API is billed per million tokens, with prices shown in "$": at off-peak times deepseek-flash costs $0.15 per million input tokens (cache miss) and $0.60 per million output tokens, while deepseek-v4-pro costs $0.66 and $1.98. Rates double during weekday peak hours (01:00–04:00 and 06:00–10:00 UTC), and cached input is much cheaper. Web search and fetch tools cost $0.75 per 1,000 successful calls. DeepSeek warns that prices may change.

Strengths and limits

• Free consumer access and low API rates, plus downloadable model weights for self-hosting.
• Privacy: DeepSeek's privacy policy says personal data is collected, processed and stored on servers in the People's Republic of China, and that user input can be used to train and improve its models; the policy describes no specific opt-out for training.
• The terms of use say the service is mainly intended for adults, with under-18s needing a guardian's consent.
• No desktop chat app is listed; DeepSeek offers a separate desktop product called DeepSeek Harness.