Gemma 3 27B IT
Neural Network
Gemma 3 27B IT is a Google multimodal model with a 128k context window for reasoning, math, and image analysis tasks.
Max answer length
(in tokens)
Context size
(in tokens)
Prompt cost
(per 1M tokens)
Answer cost
(per 1M tokens)
How it works Gemma 3 27B IT?
Gemma 3 27B IT is a Google Gemma family model: it accepts text and images as input and outputs text. The 128,000-token context window allows handling long documents and large codebases. It supports over 140 languages, with improved math, logic, and dialogue capabilities compared to the previous generation; the key difference is image support. Via BotHub, the model is available without VPN or foreign cards, with pay-as-you-go billing in rubles using Russian cards: you only pay for used tokens, which do not expire. Over 250 models are available in the same interface, allowing you to compare Gemma 3 responses with others and switch via a unified OpenAI-compatible API without rewriting integrations. Chats are encrypted with AES-GCM, and teams have access to contracts, invoices, EDI, and an admin panel with limits. The model is useful for analyzing interface screenshots or diagrams to get text descriptions, explaining and refactoring code, verifying math step-by-step, reading long reports in one context, or building a multilingual assistant via API.Frequently asked questions about Gemma 3 27B IT
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 Google model from the Gemma 3 family for text processing. It is suitable for writing and editing content, analyzing documents, explaining and fixing code, extracting data, and answering questions. It accepts both text and images/documents as input.
The single response limit is 117,964 tokens within a 128,000-token context window, shared between the prompt and the response. This is enough to generate a long article, detailed instruction, or a large block of code in one request without splitting it into parts.
A separate reasoning mode is not noted in our data for this model. However, you can achieve a similar effect via prompting: ask it to break down the task step-by-step, verify intermediate conclusions, and then formulate the final result. The context capacity is sufficient for long chains.
Yes, the model supports function calling and structured JSON output. This is convenient for agents, parsing documents into fields, and connecting external services. In BotHub, everything works via a unified OpenAI-compatible API, so you can replace the model without rewriting the integration.
You can upload images and documents: the model will analyze screenshots, diagrams, tables, or photos of text and answer based on their content. It does not accept audio or video input — for such tasks, BotHub has separate transcription and video analysis models.
The Gemma 3 family was trained on multilingual data, so the model works well with Russian queries: summarizing, rewriting, and answering questions about uploaded documents. For highly specialized terminology, you should check the quality on your specific task — in BotHub, this is available without a VPN or foreign card.