Gemini 3 Flash Preview
Neural Network
Gemini 3 Flash Preview is a fast Google model with reasoning for agent scenarios, multi-step dialogue, and coding tasks.
Specifications
- Input
- text, images, documents
- Output
- text
- Context size
- 1 048 576 tokens
- Max answer length
- 65 536 tokens
- Reasoning effort
- none, minimal, low, medium, high
- Capabilities
- reasoning, function calling
Prompt cost
(per 1M tokens)
Answer cost
(per 1M tokens)
Image prompt
(per 1K tokens)
How it works Gemini 3 Flash Preview?
Gemini 3 Flash Preview is a Google model built for speed: it reasons before answering and works confidently with tools, approaching Pro-level quality while responding significantly faster. Its profile is practical: agent scenarios, long multi-turn dialogues, and code assistance. A separate batch mode is designed for bulk processing when you have many requests and don't need an instant response. In BotHub, the model is accessible from Russia without a VPN or foreign card: you pay in rubles only for tokens used, not for a subscription, and unused tokens do not expire. Over 250 models are available in the same window, allowing you to compare answers, and you can switch to another model via a unified OpenAI-compatible API without rewriting your integration. Traffic is encrypted, and for companies, we offer contracts, invoices, electronic document management, and an admin panel with limits. Simple scenarios: build an agent that calls functions and tools; analyze code, find bugs, and suggest refactoring; maintain long support conversations without losing context; or batch-process thousands of similar tasks like labeling, field extraction, or draft descriptions.Frequently asked questions about Gemini 3 Flash Preview
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 text model with a context window of up to 1,048,576 tokens: it writes and edits text, analyzes documents, reads images, reasons before answering, and calls functions. It is suitable for analyzing large datasets, coding tasks, summarization, and scenarios requiring bulk data processing.
The model outputs up to 65,536 tokens per response — enough for a long article, detailed document analysis, or a large code file. If the material is even larger, break the task into parts and ask it to continue: the million-token context allows you to keep the entire conversation history.
Yes, the model's reasoning mode before answering is confirmed. It builds logic first and then formulates the result — this is noticeable in STEM tasks, multi-step calculations, parsing long conditions, and code debugging, where it is important not to guess the answer but to arrive at it sequentially.
Yes, function calling and structured JSON output are supported. The model can be connected to external services, search, databases, and internal tools, with responses received in a strict schema. In BotHub, this works via a unified OpenAI-compatible API, so switching models does not require rewriting your integration.
The model accepts images and documents as input — send a scan, PDF, spreadsheet, or screenshot and get a text analysis. We have no data on audio and video, so count on visual and document formats, and test other types with a specific file.
We have no specific data on language capabilities, so we cannot make promises in advance. It is more reliable to test with your own material: send a typical request, evaluate the phrasing, and switch to another model in the same window if necessary. Access from Russia without a VPN or foreign card, payment in rubles.