Grok 4.7
Neural Network
Grok 4.7 is SpaceXAI's flagship model for code and agentic tasks: connect Grok 4.7 for development in BotHub.
Specifications
- Input
- text, images, documents
- Output
- text
- Context size
- 500 000 tokens
- Max answer length
- 450 000 tokens
- Reasoning effort
- low, medium, high, extra high
- Capabilities
- function calling
Prompt cost
(per 1M tokens)
Answer cost
(per 1M tokens)
How it works Grok 4.7?
Grok 4.7 is SpaceXAI's flagship model, continuing the line after Grok 4.6, designed for programming, agentic scenarios, and knowledge work. Its strength lies in long-running engineering tasks requiring multiple consecutive steps: the model maintains the workflow and verifies its own results before providing an answer. In BotHub, you can connect to Grok 4.7 without a VPN or foreign card, pay with Russian cards based on actual token usage rather than a subscription, and switch between Grok 4.7 and other models in a single window to compare responses for the same task. For product teams, there is a unified OpenAI-compatible API: no need to rewrite integrations when switching models, chats are encrypted, and companies can work under a contract with invoicing and electronic document management. Use cases are straightforward: writing and refactoring code, debugging and completing major edits, preparing reviews and discussing architectural decisions, structuring large volumes of work materials, and building agents that execute chains of steps and verify results.Frequently asked questions about Grok 4.7
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.
Grok 4.7 is an xAI text model that works with code, long documents, and images. A 500,000-token context window allows you to keep a large project or a voluminous report entirely within the dialogue. It is suitable for analysis, refactoring, writing texts, and integration into your own services via API.
The output limit is 450,000 tokens per response, which is enough for voluminous documentation, a large code module, or a detailed analysis of material. In practice, the length depends on your request: if you need the maximum, explicitly ask for detailed text and specify the structure.
A separate reasoning mode is not noted in our data, but the model supports effort control: you can ask it to think step-by-step and show the solution process. This noticeably improves results for math, logic, and code debugging tasks.
Yes, function calling and structured JSON output are supported. The model selects the appropriate tool, inserts arguments, and returns the result in the specified schema. This is convenient for agents, data parsing, and integrations where the response goes directly into your system without manual processing.
Text, images, and documents are accepted as input, with documents processed natively without separate recognition. You can send an interface screenshot, diagram, chart, or a large PDF report and ask to analyze the content. Audio and video input is not noted in our data.
The working language is not separately described in the model's data, so it is most reliable to test it on your specific task — it will take a couple of minutes. The BotHub interface is in Russian, access is open from Russia without a VPN or foreign card, and payment is processed with Russian cards in rubles.