Gemini 2.5 Flash Lite Preview 09-2025
Gemini 2.5 Flash Lite Preview 09-2025 is Google's September 2025 preview of the next Flash Lite generation, delivering better instruction following, up to 50% fewer output tokens, and improved multimodal understanding including audio transcription and image analysis.
import { streamText } from 'ai'
const result = streamText({ model: 'google/gemini-2.5-flash-lite-preview-09-2025', prompt: 'Why is the sky blue?'})Playground
Try out Gemini 2.5 Flash Lite Preview 09-2025 by Google. Usage is billed to your team at API rates. Free users (those who haven't made a payment) get $5 of credits every 30 days.
About Gemini 2.5 Flash Lite Preview 09-2025
Gemini 2.5 Flash Lite Preview 09-2025 is a preview release from Google dated September 25, 2025. It gives you early access to the next Flash Lite tier. The preview builds on the stable Gemini 2.5 Flash Lite with three focused improvements.
First, instruction following. The preview handles complex instructions and system prompts more reliably. You'll see a smaller gap between what you ask for and what you get back. Second, verbosity. Google reported up to a 50% reduction in output tokens compared to the current stable Flash Lite. Fewer tokens means lower cost and faster responses for the same task. Third, multimodal capabilities. Audio transcription, image understanding, and translation all improved.
This preview collects developer feedback rather than replacing the stable Flash Lite. Google introduced a -latest alias system (e.g., gemini-flash-lite-latest) alongside these previews. These aliases give you automatic access to the newest version with a two-week deprecation notice. Preview models rotate. Pin to the explicit model string gemini-2.5-flash-lite-preview-09-2025 if you need consistent behavior during evaluation.
Providers
Route requests across multiple providers. Copy a provider slug to set your preference. Visit the docs for more info. Using a provider means you agree to their terms, listed under Legal.
P50 throughput on live AI Gateway traffic, in tokens per second (TPS). Visit the docs for more info.
P50 time to first token (TTFT) on live AI Gateway traffic, in milliseconds. View the docs for more info.
Direct request success rate on AI Gateway and per-provider. Visit the docs for more info.
More models by Google
| Model |
|---|
| 1M | 0.9s | 247tps | $0.25/M | $1.50/M | Read:$0.03/M Write:— | $14.00/K + input costs | — |   | 03/03/2026 | ||||
| 1M | 3.1s | 202tps | $2.00/M | $12.00/M | Read: $0.2/M Write: — | $14.00/K + input costs | — | | 02/19/2026 | ||||
| 1M | 0.7s | 190tps | $0.50/M | $3.00/M | Read: $0.05/M Write: — | $14.00/K + input costs | — | | 12/17/2025 | ||||
| 1M | 0.4s | 252tps | $0.10/M | $0.40/M | Read:$0.01/M Write:— | $35.00/K + input costs | — | | 06/17/2025 | ||||
| 1M | 0.4s | 218tps | $0.30/M | $2.50/M | Read:$0.03/M Write:— | $35.00/K + input costs | — | | 03/20/2025 | ||||
| 1M | 1.9s | 134tps | $1.25/M | $10.00/M | Read: $0.13/M Write: — | $35.00/K + input costs | — | | 03/20/2025 |
What To Consider When Choosing a Provider
- Configuration: This is a preview model. Behavior may change, or Google may deprecate it with two weeks' notice. Pin to the explicit model identifier in production and monitor for deprecation announcements.
- Zero Data Retention: AI Gateway supports Zero Data Retention for this model via direct gateway requests (BYOK is not included). To configure this, check the documentation.
- Authentication: AI Gateway authenticates requests using an API key or OIDC token. You do not need to manage provider credentials directly.
When to Use Gemini 2.5 Flash Lite Preview 09-2025
Best For
- Cost-sensitive pipelines: A 50% reduction in output tokens directly translates to lower spend at high volume
- Audio transcription and summarization: Improved multimodal handling produces more accurate text from audio inputs
- Image understanding tasks: Benefit from the enhanced visual analysis in this preview
- Multilingual translation workloads: Improved translation capabilities reduce post-processing
- System prompt-heavy applications: Rely on precise instruction following for structured output
Consider Alternatives When
- Production stability required: Pin to the stable Gemini 2.5 Flash Lite instead of a preview release
- Deep reasoning tasks: Your task requires chain-of-thought thinking, which Gemini 2.5 Flash or 2.5 Pro fits better
- Native image or audio output: Flash Lite produces text output only
- Configurable thinking budgets: A 2.5 Flash feature, not available in Flash Lite
Conclusion
This preview shows where Google is taking the Flash Lite tier: tighter instruction following, less output verbosity, and stronger multimodal input handling. Evaluate it against the stable Flash Lite to decide whether the improvements justify using a preview model in your pipeline.
Frequently Asked Questions
What improved in Gemini 2.5 Flash Lite Preview 09-2025 compared to the stable Flash Lite?
Three areas: instruction following for complex prompts, output verbosity (up to 50% fewer tokens), and multimodal capabilities including audio transcription, image understanding, and translation.
Is Gemini 2.5 Flash Lite Preview 09-2025 a stable release?
No. It's a preview release for developer feedback. Google provides a two-week deprecation notice before rotating preview models. Pin to the explicit model string if you need consistent behavior.
How much does the reduced verbosity save on cost?
Rates are listed on this page. They reflect the providers routing through AI Gateway and shift when providers update their pricing.
Does Gemini 2.5 Flash Lite Preview 09-2025 generate images or audio?
No. Like the stable Flash Lite, this model accepts multimodal inputs (text, images, audio, documents) but produces text output only.
How do I authenticate requests to Gemini 2.5 Flash Lite Preview 09-2025 through AI Gateway?
Use a Vercel API key or OIDC token with AI Gateway. Use the identifier
google/gemini-2.5-flash-lite-preview-09-2025in your requests. AI Gateway handles provider routing and failover.What is the
-latestalias system?Google introduced aliases like
gemini-flash-lite-latestthat automatically point to the newest preview. These rotate with two-week deprecation notices. Use explicit model strings for reproducibility.Should I migrate from stable Flash Lite to this preview?
Evaluate it in a staging environment first. The preview improves instruction following and reduces token usage, but behavior may change before it reaches stable. Use AI Gateway's observability to compare quality and cost side by side.