Gemini 3.5 Flash-Lite
Google DeepMind · Multimodal model
Gemini 3.5 Flash-Lite is Google's low-cost, low-latency Gemini model, generally available since July 21, 2026. It accepts the same mix of inputs as the bigger models with a 1M-token window. It suits bulk processing where price and speed matter most.
Current
ReleasedJul 21, 2026
Context window1,048,576 input tokens; 65,536 output tokens
Input → outputtext, image, video, audio, pdf → text
API price$0.30 / 1M input tokens (text/image/video/audio); $2.50 / 1M output tokens. Free tier available.
Consumer access—
Best for
High-volume, low-latency jobs such as document parsing, classification and sub-agent tasks.
Overview
Gemini 3.5 Flash-Lite is Google's low-cost, low-latency Gemini model, generally available since July 21, 2026. It accepts the same mix of inputs as the bigger models with a 1M-token window. It suits bulk processing where price and speed matter most.
Strengths and limits
- Cheapest current-generation Gemini text model
- Same 1M-token multimodal input as larger models
- Google's recommended default for new projects alongside 3.8 Flash
- Less capable than Flash/Argon on hard reasoning and coding
- Text output only