DeepSeek
Pricing checked In our news
Large language models from Chinese AI lab DeepSeek, available through a chat app and a low-cost developer API with a 1 million token context window.
No reviews yet
The score shows from 3 reviews.
Rate DeepSeek
Tap a star. No account needed.
- Category
- Models
- Made by
- DeepSeek
- Pricing
- API pay per token, from $0.15 per 1M input tokens off-peakFrom the official pricing page, checked September 25, 2026. Prices can change.
- Available on
- WebAPI
- Website
- www.deepseek.com ↗
Overview
DeepSeek is an AI lab that offers its language models through a web chat, an app download and a paid developer API. The API currently lists two models, a Flash model and a Pro model, both with a 1 million token context window. Billing is per token, and off-peak hours are charged at half the peak rate. It is aimed at everyday users who want a chatbot and at developers building applications.
Checked October 19, 2026 from deepseek.com, api-docs.deepseek.com. Something off? Tell us.
Key features
- Chat on the DeepSeek website and a downloadable app
- API with 1M token context and up to 384K output tokens
- Function calling and JSON output on both API models
- Thinking and non-thinking modes, with thinking as the default
- Image input on the Flash model
Pros
- Pay per token API with off-peak rates at half the peak price
- Large context window of 1M tokens
- Two model tiers to trade cost against capability
Cons
- Peak hours carry double the off-peak API rate
- The Pro model does not support vision input
- Concurrency is capped, at 500 for Pro and 2,500 for Flash
DeepSeek in our news
Reader reviews (0)
Used DeepSeek? Share your experience
No account needed. Takes about 10 seconds.
DeepSeek FAQ
Quick answers, plain wordsIs DeepSeek free to use?
The API is billed per token, starting from $0.15 per 1M cache-miss input tokens off-peak on the Flash model. Check the official pricing page for current chat access and rates.
What models does the DeepSeek API offer?
The pricing page lists deepseek-flash and deepseek-v4-pro. Both support a 1M token context, function calling and JSON output.
What are peak and off-peak hours?
Peak hours are 01:00 to 04:00 and 06:00 to 10:00 UTC on weekdays, excluding Chinese public holidays. All other times are off-peak and billed at half the peak rate.
Similar Models tools
View all →Launching an AI tool?
Send us the announcement. Sites showing our badge get priority review.