Google Unveils Gemini AI Family, But Delays Most Powerful Model
Google introduces Gemini as a tiered family of AI models, offering lightweight Pro and Nano versions now while delaying the flagship Ultra model until next year. The initial rollout brings Gemini Pro to Bard and enterprise tools, with a mobile-focused Nano version arriving on the Pixel 8 Pro.
Google officially unveils Gemini, its highly anticipated next-generation generative AI, but the initial release is not exactly what observers expect. Instead of launching a single powerhouse model, Google reveals Gemini 1.0 as a family of AI systems divided into three distinct tiers: the flagship Gemini Ultra, the mid-tier Gemini Pro, and the lightweight Gemini Nano designed for mobile devices.
The most capable model, Gemini Ultra, remains unavailable and is not scheduled to arrive until sometime next year. In the meantime, users experience Gemini Pro through an updated version of Google's Bard chatbot, which now uses a fine-tuned Pro model for English text interactions in the U.S. Google claims this upgrade delivers better reasoning and planning capabilities, though independent testing is not yet possible since the company skips live demonstrations.
Beyond Bard, Gemini Pro is rolling out to enterprise customers through Vertex AI and will eventually integrate into various Google products like Chrome, Ads, and Search. Meanwhile, the smallest model, Gemini Nano, launches in a limited preview exclusively for the Pixel 8 Pro through the AI Core app, allowing on-device processing for specific smartphone features before expanding to other Android devices.