Google has introduced its next-generation AI model, Gemini 1.5, which boasts enhanced performance and a breakthrough in long-context understanding. Gemini 1.5 features a new Mixture-of-Experts (MoE) architecture that improves efficiency and allows for more advanced versions of the model to be developed quickly. It offers a context window capacity of up to 1 million tokens, enabling the processing of large amounts of information. Gemini 1.5 is capable of complex reasoning, understanding and reasoning across modalities, and problem-solving tasks. The model outperforms previous versions in a range of evaluations and undergoes extensive ethics and safety testing. A limited preview is available to developers and enterprise customers, with wider release and pricing tiers coming soon.
https://blog.google/technology/ai/google-gemini-next-generation-model-february-2024/