Google launched Gemini 3.1 Flash-Lite, the fastest model in the Gemini 3 family, on Tuesday.
The model targets high-volume developer workloads at a competitive price point, positioning it against rivals like GPT-5 mini and Claude 4.5 Haiku. Google offers the model at $0.25 per million input tokens and $1.50 per million output tokens, making it the most affordable option in the Gemini 3 lineup. The company stated the model is available in preview via the Gemini API in Google AI Studio and Vertex AI, but it is not available in the consumer app.
Benchmarks indicate the model generally outperforms Gemini 2.5 Flash at a lower price. Google promises a generation speed of up to 363 tokens per second, which is two to five times faster than competitors. However, the company noted this is three tokens slower than the previous Gemini 2.5 Flash-Lite model.
Video: Google
On the Arena.ai Leaderboard, Gemini 3.1 Flash-Lite scores 1432 Elo points. This score places it among open-weight models and last-generation commercial offerings. Google did not publish agent benchmarks, stating the model is intended for data processing rather than managing fleets of agents.
Developers can use the API to adjust the model’s reasoning time to balance cost and performance. This option allows users to minimize token production for high-volume tasks. The launch deviates from Google’s usual pattern of releasing a more capable Flash version first.
