DeepSeek R1
Neural Network
DeepSeek R1 is an open reasoning model from DeepSeek: suitable for complex logic, STEM tasks, and coding.
Max answer length
(in tokens)
Context size
(in tokens)
Prompt cost
(per 1M tokens)
Answer cost
(per 1M tokens)
How it works DeepSeek R1?
DeepSeek R1 is an open-source reasoning model from DeepSeek: 671 billion parameters, with 37 billion active per token. Reasoning tokens are visible, so you see the thought process, not just the final answer. In BotHub, it is available from Russia without a VPN or foreign card: pay with Russian cards as you go for used tokens, which do not expire. With over 250 neural networks in one window, it's easy to compare DeepSeek R1 with other models on the same task. Use a unified API to connect DeepSeek to your services and switch models without rewriting integrations. Data is encrypted via AES-GCM and not stored. Companies get contracts, invoices, EDI, and an admin panel with limits. Useful for debugging code, refactoring, code reviews, solving step-by-step math or STEM problems, verifying analytical logic, or building a multi-model service where R1 handles complex reasoning.Frequently asked questions about DeepSeek R1
You can use generated results for commercial purposes. You own all rights to the content you create. The only restriction: make sure your prompt does not include copyrighted third-party material. You are responsible for respecting the rights to any input data.
deepseek-r1 is a model that reasons before answering: suitable for code analysis and debugging, math and STEM tasks, document and long text analysis. It accepts text, documents, and images as input and supports function calling, making it suitable for analytics and development workflows.
The maximum response is 16,000 tokens, and the total context window is 64,000 tokens, so a large document with a detailed analysis fits in one request. If there is more material, break the task into parts and ask it to continue from the right place.
Yes, this is its main mode: deepseek-r1 builds a chain of thought and only then provides the final answer. Therefore, the model handles multi-step math, logic, and code analysis more accurately — it verifies intermediate conclusions itself. Reasoning also consumes tokens.
Yes. Function calling and structured JSON output are supported, so the model can be connected to tools, databases, and external services. In BotHub, this works via a unified OpenAI-compatible API: no need to rewrite integrations, just change the model identifier in the request.
The model accepts images and documents as input — you can send a screenshot, diagram, or file and ask it to analyze the content. There is no mention of audio or video in our data. deepseek-r1 responds with text, so for audio or image generation, switch to a specialized model in the same window.
Our data does not specify language performance, so we won't make promises — it's easier to test with your own query. The service itself is Russian-speaking: interface, support, and payment with Russian cards in rubles, without VPN or foreign cards.