DeepSeek is a Chinese AI company that develops high-performance large language models at a fraction of the cost of competitors. Its flagship DeepSeek-V3.2 model powers both the free web chatbot and the developer API, offering two modes: deepseek-chat for general conversation and deepseek-reasoner for advanced step-by-step reasoning. DeepSeek gained global attention for training its models at remarkably low costs, with the V3 model reportedly trained for just $5.9 million, challenging the assumption that frontier AI requires billions in compute. The API features automatic context caching that reduces costs by up to 90% on repeated prompts, with cache hit pricing at just $0.028 per million tokens. The web app at chat.deepseek.com is entirely free to use with no usage limits, while the API offers some of the most affordable token pricing in the industry. DeepSeek supports 128K context windows, JSON output mode, tool calling, and function calling. The platform has become especially popular among developers and researchers seeking high-quality AI capabilities at accessible price points.