
Google on Tuesday unveiled significant updates to its Gemini 2.5 artificial intelligence model lineup, introducing a new budget-friendly option and stabilizing pricing across its flagship AI offerings as demand continues to surge.
New Flash-Lite Model Targets Cost-Conscious Developers
The tech giant’s latest addition, Gemini 2.5 Flash-Lite, represents the most affordable and fastest option in the company’s 2.5 model series. The new model is designed as an upgrade from Google’s previous 1.5 and 2.0 Flash versions, delivering improved performance while reducing both response time and operational costs.

“Flash-Lite is optimized for high-throughput applications like large-scale classification and summarization tasks,” Google said in a blog post announcing the updates.
Unlike other models in the 2.5 family, Flash-Lite ships with its “thinking” capability disabled by default to maximize speed and minimize costs. Developers can activate the reasoning feature through an API parameter when needed.
Pricing Changes Reflect Strong Performance
Google also announced pricing adjustments for its standard Gemini 2.5 Flash model, citing the system’s “exceptional value” and performance improvements. The company described the model as offering the “best cost-per-intelligence available” in the current market.
The stable version of Gemini 2.5 Flash, which matches the preview model showcased at Google’s I/O conference in May, is now generally available. Users of the older 04-17 preview version have until July 15, 2025, before that model endpoint shuts down.

Pro Model Sees Record Demand
Google reported that its premium Gemini 2.5 Pro model is experiencing unprecedented adoption rates, with the company calling demand growth “the steepest of any of our models we have ever seen.”
To accommodate increased usage, Google has made the 06-05 version of Pro stable while maintaining existing pricing. The model targets high-complexity applications including software development and autonomous agent tasks.

“We expect that cases where you need the highest intelligence and most capabilities are where you will see Pro shine, like coding and agentic tasks,” the company stated.
Users of the 05-06 Pro preview have until June 19, 2025, to migrate to newer versions before service discontinuation.
Reasoning Technology at Core
All Gemini 2.5 models feature “thinking” capabilities that allow the AI systems to process problems internally before generating responses. This approach aims to improve accuracy and performance compared to traditional large language models.

Developers can adjust the “thinking budget” to balance computational costs against response quality, giving them granular control over AI behavior based on specific use cases.
The updates come as Google intensifies competition with rivals including OpenAI and Anthropic in the rapidly evolving generative AI market. The company’s focus on cost optimization and performance scaling reflects industry-wide pressure to make advanced AI capabilities more accessible to developers and businesses.
Google indicated additional announcements regarding model scaling beyond the Pro tier are forthcoming.