Gemini 3 Pro and Flash: Google relaunches the frontier challenge
In one sentence Google DeepMind announces Gemini 3 with Pro and Flash variants: improved reasoning, native long context, deeper integration into Workspace and Android.
Google updates its Gemini model family to version 3, after a 2025 in which Gemini 2.5 Pro gained ground against GPT-5 and Claude 4 but never quite took clear leadership.
Gemini 3 ships in two main variants: Pro (complex tasks, reasoning, code) and Flash (faster and cheaper, aimed at agents and high-volume apps). Both keep Google's signature traits: huge context window and native multimodal handling (text, images, audio, video).
For Google users: deeper presence in Workspace (Gmail, Docs, Sheets), the Android assistant, and Search via AI Overviews. For developers: AI Studio and Vertex AI updates with competitive pricing for the Flash tier.
Not a revolutionary leap like Gemini 1.5 was for long context, but a clear signal that Google stays in serious contention with OpenAI and Anthropic on frontier models.
Companies
Google, Google DeepMind
Tools
Gemini 3 Pro, Gemini 3 Flash
Tags
Sources