Grok 4.3
Neural Network
Grok 4.3 is a reasoning model from xAI for agentic scenarios and text/image processing, available via Grok 4.3 API.
Specifications
- Input
- text, images, documents
- Output
- text
- Context size
- 1 000 000 tokens
- Max answer length
- 900 000 tokens
- Reasoning effort
- none, low, medium, high
- Capabilities
- function calling
Prompt cost
(per 1M tokens)
Answer cost
(per 1M tokens)
How it works Grok 4.3?
Grok 4.3 is a reasoning model from xAI that accepts text and images as input and outputs text. It is designed for agentic scenarios, precise instruction following, and tasks requiring high factual accuracy. Before answering, it builds a chain of reasoning, making it reliable for multi-step tasks: parsing conditions, planning actions, coding, and data analysis. In BotHub, the model works without a VPN or foreign card: pay with a Russian card as you go, per token, without a subscription, and unused tokens do not expire. Over 250 other neural networks are available in the same window for easy comparison and switching. Developers get a unified OpenAI-compatible API, so no integration rewrite is needed; chats are encrypted with AES-GCM. Companies have access to contracts, invoices, electronic document management, and an admin panel with employee limits. Scenarios: agentic pipelines and multi-step process automation, parsing screenshots, diagrams, and charts, writing and debugging code, checking large instructions and regulations, and preparing fact-based analytical reports.Frequently asked questions about Grok 4.3
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 xAI with a million-token context. It accepts text, images, and documents as input, can call external functions, and return structured JSON. Suitable for parsing large volumes of documentation, coding, analytical, and research tasks.
Up to 900,000 tokens in a single response — enough for voluminous texts: large reports, documentation, transcripts, or an entire codebase. The actual length depends on the task and prompt, but you will rarely hit the output limit. Pay only for the tokens actually used.
Our data doesn't explicitly mark a reasoning mode, but that doesn't mean it's absent. The model has an 'effort' parameter to adjust the depth of the response. For complex tasks, simply ask it to break down the solution step-by-step.
Yes, both. The model supports function calling and structured JSON output, making it easy to integrate into agents, chatbots, and data processing pipelines. In BotHub, this works via a unified OpenAI-compatible API — you can switch models without rewriting your integration.
According to our data, it accepts text, images, and documents as input, including native formats: the model can parse screenshots, diagrams, tables, or multi-page files and respond with text. We have no information regarding audio or video. The output is text.
We have no information about prompt and response languages, so it's safer to test the model directly on your task — it's quick. The BotHub interface and ruble payments are fully Russian: access without VPN or foreign cards, with other models available in the same window for comparison.