Vai al contenuto
AI.info

tools

DeepSeek - Open-Source Reasoning AI | AI.info

DeepSeek R1 and V3 are frontier open-source AI models matching GPT-4 class performance. Free web interface plus low-cost API access.

DeepSeek - Open-Source Reasoning AI | AI.info

In inglese

DeepSeek provides web and mobile chat with writing help, coding assistance, question answering, web search, Deep-Think reasoning, file uploads, text extraction, and synchronized chat history. Its current services also support multimodal understanding and long-context conversations.

The chat service is free to use. API access is billed by token usage, with separate peak and off-peak rates. The website could not be opened during this search because it returned HTTP 403.

Features

  • Answer questions and assist with writing, coding, research, and explanations
  • Search the web for information
  • Use Deep-Think reasoning mode
  • Upload files and extract text
  • Support native visual understanding
  • Support up to 1M tokens of context
  • Provide API access with tool calls, JSON output, and Responses API support
  • Sync chat history across supported platforms

Use cases

  • Draft and revise emails, documents, and other written content
  • Explain code and help solve programming problems
  • Research topics and summarize web information
  • Analyze uploaded files and images
  • Work through difficult reasoning and mathematics problems
  • Integrate AI responses into applications through the API

Pros

    Cons

      Latest updates

      • DeepSeek-V4.1-Flash Release (DeepSeek-V4.1-Flash)

        DeepSeek V4.1 Flash is available on the API with native multimodal support; legacy model names route to V4.1 Flash.

      • DeepSeek-V4-Flash-Vision-Exp Release (DeepSeek-V4-Flash-Vision-Exp)

        The model is available on the DeepSeek API platform and can be accessed by setting model='deepseek-v4-flash-vision-exp'.

      • DeepSeek-V4-Pro Update (DeepSeek-V4-Pro)

        The GA release rolled out on the APP, Web, and API; the API supports the Responses API format; thinking modes support low / high / max.

      • DeepSeek-V4-Flash Update (DeepSeek-V4-Flash)

        The DeepSeek-V4-Flash API is in public beta; it supports the Responses API format and is adapted for Codex.

      • DeepSeek-V4 (DeepSeek-V4)

        The API supports V4-Pro and V4-Flash through the OpenAI ChatCompletions interface and the Anthropic interface.

      Capabilities

      • Searches the web — “🔍 Web search & Deep-Think mode” source
      • Reads your files — “📄 File upload & text extraction” source
      • Understands images — “The official version of DeepSeek-V4.1-Flash has been released, featuring comprehensive improvements in text and Agent performance, along with native multimodal visual understanding capabilities.” source
      • Developer API — “BASE URL (OpenAI Format) https://api.deepseek.com” source
      • Mobile app — “📱 Now officially available on App Store & Google Play & Major Android markets” source

      Pricing

      Prices checked
      2026-09-25

      deepseek-flash

      • $0.003 per 1M input tokens (cache hit), off-peak
      • $0.006 per 1M input tokens (cache hit), peak
      • $0.15 per 1M input tokens (cache miss), off-peak
      • $0.3 per 1M input tokens (cache miss), peak
      • $0.6 per 1M output tokens, off-peak
      • $1.2 per 1M output tokens, peak
      • Context length: 1M
      • Maximum output: 384K
      • Supports non-thinking and thinking modes
      • JSON Output
      • Tool Calls
      • Vision supported

      deepseek-v4-pro

      • $0.022 per 1M input tokens (cache hit), off-peak
      • $0.044 per 1M input tokens (cache hit), peak
      • $0.66 per 1M input tokens (cache miss), off-peak
      • $1.32 per 1M input tokens (cache miss), peak
      • $1.98 per 1M output tokens, off-peak
      • $3.96 per 1M output tokens, peak
      • Context length: 1M
      • Maximum output: 384K
      • Supports non-thinking and thinking modes
      • JSON Output
      • Tool Calls
      • Vision not supported
      Official website