Gemini 3.5 Flash-Lite

Gemini 3.5 Flash-Lite is the latest in our cost-effective Flash-Lite line of models. It's optimized for simple coding tasks, precise document understanding, and lightweight agentic workflows that require fast inference at minimal cost. 3.5 Flash-Lite is a suitable replacement model for Gemini 2.5 Flash and less complex Gemini 3 Flash workloads.

When using 3.5 Flash-Lite, you can use different thinking levels to optimize for quality, speed, and token output:

• Use thinking_level.MINIMAL for latency-sensitive or simpler classification and extraction tasks: 3.5 Flash-Lite defaults to minimal thinking, which optimizes for speed and cost efficiency. This is ideal for high-throughput classification, routing, or JSON extraction tasks.
• Use thinking_level.MEDIUM or thinking_level.HIGH for subagents: If you're using 3.5 Flash-Lite as an autonomous subagent that writes code, runs terminal commands, or calls external APIs, set the thinking level to medium or high. Using the minimal thinking level can cause premature tool termination on multi-step tasks.

3.5 Flash-Lite includes the following potentially breaking changes when compared to previous Gemini models:

• Custom values for parameters like temperature, top-K, and top-P aren't supported. If you set a custom value for these parameters, that value will be ignored.
• Custom values for frequency and presence penalty parameters aren't supported. Setting a custom value for these parameters will throw an error.
• API requests where the last input turn has a role of Model aren't supported. The following kinds of requests will return an error:
  • When using the Interactions API: Requests where the last object in the input array has "type": "model_output".
  • When using the GenerateContent API: Requests where the last object in the contents array has "role": "model".

Try in Agent Studio Deploy example app View pricing

Note: "Deploy example app" requires a Google Cloud project with billing and Agent Platform API enabled.

Model ID	gemini-3.5-flash-lite
Modalities	description Text Input and output photo Image Input only mic Audio Input only videocam Video Input only
Token limits	Context window	1,048,576
	Maximum output tokens	65,536
Capabilities	Thinking Supported System instructions Supported Gemini Live API Not supported Structured output Supported Context caching Implicit context caching, explicit context caching Supported Count Tokens Supported RAG Engine Supported Chat completions Supported Tuning Not supported URL context Supported
Tools	Grounding Google Search, Google Maps Supported Code execution Supported Function calling Supported Computer Use Preview feature Not supported
Consumption options	Provisioned Throughput Supported Batch inference Supported Pay-as-you-go Standard PayGo, Flex PayGo, Priority PayGo Supported Fixed quota Not supported
Technical specifications	Image	Maximum images per prompt: 3,000 Maximum file size per file for inline data or direct uploads through the console: 7 MB Maximum file size per file from Google Cloud Storage: 30 MB Maximum number of output images per prompt: 10 Supported MIME types: image/png, image/jpeg, image/webp, image/heic, image/heif
	Text	Maximum number of files per prompt: 3,000 Maximum number of pages per file: 3,000 Maximum file size per file for the API or Cloud Storage imports: 50 MB(application/pdf) or 7 MB(text/plain) Maximum file size per file for direct uploads through the console: 7 MB Supported MIME types: application/pdf, text/plain
	Video	Maximum video length (with audio): Approximately 45 minutes Maximum video length (without audio): Approximately 1 hour Maximum number of videos per prompt: 10 Supported MIME types: video/x-flv, video/quicktime, video/mpeg, video/mpegs, video/mpg, video/mp4, video/webm, video/wmv, video/3gpp
	Audio	Maximum audio length per prompt: Approximately 8.4 hours, or up to 1 million tokens Maximum number of audio files per prompt: 1 Supported MIME types: audio/x-aac, audio/flac, audio/mp3, audio/m4a, audio/mpeg, audio/mpga, audio/mp4, audio/ogg, audio/pcm, audio/wav, audio/webm
	Parameter defaults	Temperature: 1.0 topP: 0.95 topK: 64 Frequency penalty: 0 Presence penalty: 0
Supported regions	Model availability	Global: global Multi-region: us, eu
	ML processing	Multi-region: us, eu
	Provisioned Throughput	Global: global Multi-region: us, eu
	Standard PayGo	Global: global Multi-region: us, eu
Versions	gemini-3.5-flash-lite Launch stage: GA Release date: July 21, 2026 Retirement date: July 21, 2027 or later
Security controls	Online prediction	Data residency CMEK VPC-SC AXT
	Batch inference	Data residency CMEK VPC-SC AXT
	Context caching	Data residency CMEK VPC-SC AXT
	See Security controls for more information.