Kimi K3
Neural Network
Kimi K3 is an open-weights multimodal reasoning model from Moonshot AI, suitable for complex code and agent scenarios.
Max answer length
(in tokens)
Context size
(in tokens)
Prompt cost
(per 1M tokens)
Answer cost
(per 1M tokens)
How it works Kimi K3?
Kimi K3 is an open-weights multimodal reasoning model from Moonshot AI with 2.8 trillion parameters. It excels at complex programming, knowledge work, and long agent scenarios where tasks are solved via step-by-step reasoning rather than a single answer. The batch version is designed for bulk request processing. Via BotHub, the model is accessible without a VPN or foreign card: pay with a Russian card in rubles only for tokens used, with no subscription. Access over 250 models in one window to switch and compare responses on the same prompt. A unified OpenAI-compatible API lets you change models in your service without rewriting integrations. Chats are encrypted, and companies get contracts, invoices, and an admin panel with limits. Developers use Kimi K3 for refactoring, analyzing unfamiliar codebases, and debugging. Analysts use it to verify facts from large materials, product teams build agents that break down tasks, and researchers use it to analyze technical calculations.Frequently asked questions about Kimi K3
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.
This is a text model from Moonshot AI with a 1,048,576 token context. It accepts text, documents, and images, reasons before answering, and calls functions. It is suitable for analyzing large documentation, working with code (writing, refactoring, debugging), as well as for analytics and STEM tasks.
The model outputs up to 16,384 tokens per response—enough for a large documentation section, a major code module, or a detailed analysis. Need more? Break the task into parts: the 1,048,576 token context window lets you keep the entire dialogue and source files in view.
Yes, reasoning before answering is confirmed: the model first builds a chain of steps and only then formulates the result. This significantly helps with math, logic, resolving conflicting requirements, and debugging, where it is important to reach the answer sequentially rather than guessing.
Yes, function calling and structured JSON output are supported, so the model can be connected to tools, databases, and internal services. In BotHub, everything goes through a unified OpenAI-compatible API: you can switch to another model without rewriting integrations, and tokens do not expire.
Images and documents—yes: send a screenshot, diagram, PDF, or table and get a text analysis. Audio and video are not listed in our data, so do not include them in your scenario; if necessary, switch to a specialized model in the same window.
We do not have specific language performance data, so we cannot make promises—test the model with your own prompts and compare it with others in the same window. The website and Telegram bot interface are in Russian, payment is via Russian cards, and access is available without a VPN or foreign card.