Low-cost AI models with a chat app and a very cheap API for developers.
- Company
- DeepSeek
- Best for
- Developers and businesses that want strong reasoning and coding at a low API price
- Free plan
- Chat app is available to use; plan prices are not listed on the pages we checked
- Starting price
- API from $0.15 per 1M input tokens (off-peak, US dollars). No subscription prices listed on the pages we checked
- Details checked
- 28 October 2026
Live numbers
- Public interest (28 days)
- 89,638 views Steady
Updated 28 October 2026, 15:38. See the full dashboard.
Standout features
- Very low API prices, with cached input as low as $0.003 per 1M tokens
- Half-price off-peak rates, including weekends
- 1M-token context window and thinking mode on both API models
- Works with OpenAI-format and Anthropic-format API calls, so existing tools can switch easily
Best uses
- Coding help and code review
- Maths, logic and step-by-step reasoning
- Bulk text jobs run cheaply through the API
- Powering your own chatbot or automation at low cost
Pros
- Among the cheapest options for developers building on an API
- Strong at reasoning, maths and coding tasks
- Simple pricing page with no tiers to compare
- Easy to plug into tools that already use the OpenAI or Anthropic API format
Cons
- No UK or GBP pricing: everything is in US dollars
- Peak-hour prices are double, so cost depends on when you run jobs
- Data is processed by a China-based company; check this against your privacy and client requirements before sending business or personal data
- Fewer built-in business features (teams, admin controls) than the big assistants
How to use DeepSeek
- Go to deepseek.com and open the chat app, or create an API account at platform.deepseek.com if you are a developer.
- Start with a clear task and say who the answer is for.
- For hard problems, use thinking mode and ask it to show its working.
- For the API, use the model name deepseek-flash and point your OpenAI-style client at https://api.deepseek.com.
- Do not paste customer or personal data until you have checked the privacy terms.
Prompts that work well
Reason first
Work through this step by step, then give a final answer in one sentence: [problem]
Code review
Review this code for bugs and security problems. List each issue with a fix. [paste code]
Compare options
Compare these two approaches in a table: cost, effort, risk. Then recommend one for a small business. [options]
Pricing
| Plan | Price | What you get |
|---|---|---|
| deepseek-flash (API) | $0.15 in / $0.60 out per 1M tokens off-peak | Cache-miss input price. Peak hours cost double. Supports thinking and non-thinking modes, vision, 1M context. |
| deepseek-v4-pro (API) | $0.66 in / $1.98 out per 1M tokens off-peak | Cache-miss input price. Peak hours ($1.32 in / $3.96 out) cost double. 1M context, no vision. |
| Peak hours | 2x price | 01:00-04:00 and 06:00-10:00 UTC, Monday to Friday, excluding Chinese public holidays. Prices are in US dollars, per DeepSeek API docs. |
Prices change often. Always confirm on the official site. We last checked on 28 October 2026.
Similar tools
Chatbots and assistantsClaude1.0% share Steady interest +15%AI assistant for writing, long documents, coding and analysis, with a free plan.
Chatbots and assistantsChatGPT80.5% share ▲ Growing interest +18%General-purpose AI assistant for writing, research, coding and images, with a free plan.
Chatbots and assistantsMicrosoft Copilot1.8% share Steady interest -1%Microsoft’s AI assistant, built into Word, Excel, PowerPoint and Outlook.
Chatbots and assistantsMistralinterest +4%European AI assistant and models, with a free plan and low-cost Pro.