Gemini 3.1 Flash Lite
Neural Network
Gemini 3.1 Flash Lite is a low-latency multimodal Google model for high-volume tasks and lightweight agents; available via API in BotHub.
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, Batch API
Prompt cost
(per 1M tokens)
Answer cost
(per 1M tokens)
Image prompt
(per 1K tokens)
How it works Gemini 3.1 Flash Lite?
Gemini 3.1 Flash Lite is a Google model from the Gemini line designed for low latency and high request volume. It understands more than just text: you can input images, video, audio, and PDFs. The model is focused on lightweight agent scenarios where response speed and predictable cost per request are important. In BotHub, it is accessible without a VPN or foreign card: you pay in rubles only for the tokens used, with no subscription, and your balance does not expire. All 250+ models are gathered in one window, so you can switch to another in a few clicks, and a unified OpenAI-compatible API eliminates the need to rewrite integrations. For companies, we offer contracts, invoices, electronic document management, and an admin panel with employee limits. Simple scenarios include: mass classification of requests, data extraction from PDFs and scans, analysis of screenshots and video frames, rough code work and debugging tips, and fast chatbots and agents that need to respond instantly.Frequently asked questions about Gemini 3.1 Flash Lite
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 with a 1,048,576 token context: it analyzes large documents and codebases, answers based on file and image content, reasons before answering, and calls external functions. Suitable for analytics, support, automation, and batch processing of large volumes.
The output limit is 65,536 tokens per response. This is enough for a detailed technical analysis, a large report, or an entire code module. If the material doesn't fit, ask for a continuation in parts: the million-token context allows keeping the entire conversation in the dialogue memory.
Yes, reasoning mode is supported: the model first builds a solution path and then provides an answer. This is noticeable in tasks with nested logic, calculations, code debugging, and parsing conflicting requirements, where it is important not just to guess the result, but to trace the steps.
Yes, the model supports both function calling and structured JSON output. You can connect it to your services as a tool executor and receive a predictable response schema for parsing. In BotHub, this is done via a unified OpenAI-compatible API without rewriting integrations.
According to our data, images and documents are accepted as input, including native file processing — diagrams, screenshots, tables, PDFs. Audio and video support for this model is not noted in our data, so for such tasks, it is safer to switch to a specialized model.
We do not have specific data on languages, so we cannot promise accuracy: test the model with your typical query. It's fast — the BotHub interface is in Russian, payment via Russian cards without VPN is available, and switching between models in one window takes seconds.