Google’s Gemini 3.5 Flash-Lite supports high-volume multimodal tasks, configurable thinking, and document parsing with a 1M-token input limit.
At a glance
API model ID: gemini-3.5-flash-lite
Capabilities & boundaries
Tool support requires the appropriate API integration; a supported tool is not automatically active in every chat.
Start with minimal for bounded extraction and classification. Increase the thinking level only if tests show a quality improvement. Minimal does not guarantee zero thinking; difficult inputs can still require some reasoning.
Put it to work
Original examples from EZ Ai Assist. Adapt these to your task and the features available in your workspace.
Workflow 01
Use a supplied invoice and a strict schema; keep missing fields visible.
Extract vendor, invoice number, issue date, currency, subtotal, tax, and total from the invoice I provide. Return one JSON object and a list of fields that were unreadable or absent. Do not infer missing amounts. Check whether the supplied subtotal plus tax matches the total and report any discrepancy.
Workflow 02
Keep routing bounded to the categories you actually support.
Classify each support ticket below into billing, login, product question, or needs human review. Return the ticket ID, category, urgency, and one short evidence quote. Preserve the original IDs. If the ticket is ambiguous or asks for an account change, choose human review instead of inventing a resolution.
Workflow 03
Specify what may change and what must remain literal.
Normalize these product records into the provided schema. Preserve SKUs, prices, currency codes, and quoted dimensions exactly. Standardize only category spelling and whitespace. Return valid rows separately from rows needing review, with the reason for each rejection. Do not add product claims or fill missing specifications.
Developer reference
These are Google API reference prices, not EZ Ai Assist subscription prices.
View EZ Ai Assist plans| Token type | Price |
|---|---|
| Input | $0.30 |
| Cached input | $0.03 |
| Output (including thinking) | $2.50 |
| Usage | Rate and unit |
|---|---|
| Cache storage | $1.00per 1M tokens per hour |
| Google Search grounding (after allowance) | $14.00per 1,000 search queries |
| Google Maps grounding (after allowance) | $14.00per 1,000 search queries |
Common questions
Consider it for document parsing and lightweight agent tasks. Supply clear constraints and assess correctness, latency, and total usage on your own examples before making it a default.
This Flash-Lite model supports minimal, low, medium, and high thinking levels, with minimal as the default. More thinking can increase response time and output-token cost; it is not a substitute for validation.
The Google model card supports these inputs and text output. Upload and tool availability depend on your integration. This does not make the model an image generator, voice generator, or Live API model.
Compared with 3.1 Flash-Lite, 3.5 adds preview computer-use support and has its own token rates. Compare extraction accuracy and total latency on the same inputs; a newer version is not automatically the cheapest.
Cache storage and grounding can add charges beyond ordinary input/output tokens. Audio input has a separate rate where shown. Review the units and free allowances in the provider pricing page.
No. Keep evidence relevant, specify the required output, request source references, and check the result. The 1,048,576 input-token limit and 65,536 output-token limit are different constraints.
No. This guide separates Google’s direct API reference from EZ Ai Assist subscriptions. Use the pricing page for plans and the app for current model access. The example prompts are starting points, not guaranteed results.
Check the source
Specifications and API prices checked on . Example prompts and workflow advice are editorial guidance from EZ Ai Assist.
Same provider
Google’s Gemini 3.8 Flash supports multimodal understanding, three thinking levels, and tool-assisted workflows with promotional API pricing.
Google’s Gemini 3.7 Flash combines multimodal inputs, configurable thinking, and structured tool workflows for repeatable everyday tasks.
Google’s Gemini 3.1 Pro Preview supports complex synthesis, coding, and multimodal analysis with a 1M-token input limit.
Historical guide to Google’s retired Gemini 3 Pro Preview API, with specifications, migration checks, and the Gemini 3.1 Pro replacement.