Google Unveils Budget-Friendly Gemini 3.1 Flash-Lite AI Model
Google introduces the Gemini 3.1 Flash-Lite AI model, offering significantly lower costs and faster speeds than its predecessors. The new model outperforms competitors on several benchmarks and targets high-volume, routine developer tasks.
Google launches Gemini 3.1 Flash-Lite, a new addition to its Gemini series of multimodal artificial intelligence models designed specifically for cost-efficiency. The model costs just $0.25 per million input tokens and $1.50 per million output tokens, making it significantly cheaper than the capable Gemini 3.1 Pro. Google emphasizes that this budget-friendly pricing structure allows developers to run high-volume workloads without incurring massive expenses.
Beyond its low cost, Gemini 3.1 Flash-Lite delivers impressive performance gains over previous budget-focused iterations. The model generates answers 45% faster than Gemini 2.5 Flash and reduces the wait time for the first output token by 2.5 times. It processes multimodal prompts of up to 1 million tokens and generates up to 64,000 tokens of text, which includes the ability to create code-based visual assets like business intelligence dashboards.
The new algorithm achieves top scores in six out of 11 benchmark tests, outperforming competitors like GPT-5 mini and Anthropic's Claude 4.5 Haiku. Google envisions developers using this model for routine, high-volume tasks that do not require deep reasoning, such as translating e-commerce product listings, enforcing terms of service, and rapidly populating website prototypes with sample data. The model is currently available in preview.