Gemini 3 Flash
Overview
Gemini 3 Flash is a multimodal language model developed by Google DeepMind, released on December 17, 2025. It is part of the Gemini 3 family and is designed to deliver high-speed, efficient performance across various modalities, including text, image, audio, video, and PDF inputs. The model features a context window of up to 1 million tokens and supports outputs up to 64,000 tokens. Notably, Gemini 3 Flash is priced at $0.50 per million input tokens and $3.00 per million output tokens, positioning it as a cost-effective option for applications requiring rapid processing. In terms of performance, Gemini 3 Flash demonstrates strong capabilities in several benchmark evaluations. It achieved a score of 33.7% on Humanity’s Last Exam, a benchmark assessing expertise across various domains, and 90.4% on the GPQA Diamond benchmark, which tests scientific knowledge. Additionally, it scored 78% on the SWE-bench Verified coding benchmark, indicating robust coding proficiency. These results suggest that Gemini 3 Flash is well-suited for tasks involving complex reasoning, scientific inquiry, and software development. Its efficiency and multimodal processing capabilities make it a competitive choice among contemporary AI models. Gemini 3 Flash is available through Google’s Gemini app, AI Mode in Search, and developer tools such as Google AI Studio, Gemini CLI, and Antigravity. Enterprise users can access the model via Vertex AI and Gemini Enterprise. This accessibility, combined with its performance metrics and pricing, positions Gemini 3 Flash as a versatile and efficient solution for a wide range of AI applications.