Reka Edge
Neural Network
Reka Edge is a compact multimodal model from Reka for analyzing images and video with text responses.
Specifications
- Input
- text, images, documents
- Output
- text
- Context size
- 16 384 tokens
- Max answer length
- 14 745 tokens
- Capabilities
- function calling
Prompt cost
(per 1M tokens)
Answer cost
(per 1M tokens)
How it works Reka Edge?
Reka Edge is a compact 7B parameter multimodal model from Reka: it accepts images, video, and text, and responds with text. It is optimized primarily for understanding visual content, so the model analyzes frame contents and formulates observations in words while remaining resource-efficient. In BotHub, Reka Edge works without a VPN or foreign card: you pay with a Russian card in rubles and only for tokens used, with no subscription. Over 250 models are available in the same window, making it easy to run one task through several neural networks and compare answers, while a unified API allows replacing the model without rewriting the integration; data is protected by AES-GCM encryption, and companies have access to contracts, invoices, and an admin panel with limits. It is useful when you need to describe product photos for a catalog, analyze a user-submitted screenshot in support, extract the essence from a short video clip, label and categorize a media library, or prepare draft captions for images.Frequently asked questions about Reka Edge
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.
Reka Edge is a compact Reka model that works with text, documents, and images via BotHub. It is convenient for parsing PDFs and scans, extracting data, answering questions based on documents, and summarizing. Available without VPN or foreign cards, payment is per token used.
According to our data, the output limit for Reka Edge is 14,745 tokens per response, and the entire conversation, including the prompt and attachments, fits into a 16,384 token context. It is better to submit long documents in parts or via follow-up queries to fit within this window.
Our data does not indicate specific reasoning capabilities before answering, so do not expect long internal deliberations from Reka Edge. The model answers to the point and handles tasks well where direct analysis of a document or image is needed, rather than multi-step reasoning.
Yes, according to our data, Reka Edge supports function calling and structured JSON output. This allows it to be embedded into agent scenarios and services where the model's response must return in a specific schema rather than free text. The unified OpenAI-compatible BotHub API simplifies this integration.
Images and documents — yes, this is stated in Reka Edge's capabilities. Audio and video are not noted in our data, so do not count on them. Through BotHub, the model accepts an image or file along with a text query and answers based on their content.
There is no specific note on Russian language quality in our data, but Reka describes its models as multilingual. Practical advice: test Reka Edge with your own phrasing and documents — a couple of test queries will be enough for data extraction and answering based on files.