From the archive
Gemini 1.5 Pro opens a long-context preview
Google introduces a limited Gemini 1.5 Pro preview with up to one million tokens through AI Studio and Vertex AI.
Google DeepMind introduced Gemini 1.5 on February 15, 2024. The first model, Gemini 1.5 Pro, opened for limited testing through Google AI Studio and Google Cloud Vertex AI .
Longer material in one request
The preview included an experimental context window of up to one million tokens for a limited group of developers and enterprise customers. That increased the amount of text, audio, video, or code a request could contain. Google described a standard 128,000-token window for the planned wider release.
The announcement illustrated the longer context with tasks involving lengthy documents and video. The feature expanded input capacity rather than removing the need to evaluate how well a model used the supplied material.
Two access routes
AI Studio provided a route for testing Gemini through Google's developer tools. Vertex AI placed the model within Google Cloud's deployment environment. Wider availability and pricing tiers were still planned at this stage.