07/08/2024, 08:44 AM UTC
谷歌云更新Vertex AI平台，增强模型和企业功能Google Cloud Updates Vertex AI Platform with Enhanced Models and Enterprise Features
1. 谷歌云正式发布Gemini 1.5 Flash，提供低延迟、有竞争力的价格和突破性的100万个令牌上下文窗口，适用于从零售聊天代理到文档处理和研究代理等广泛的大规模用例。 2. Gemini 1.5 Flash在GPT-3.5 Turbo等同类模型中提供显著优势，拥有60倍更长的上下文窗口，对于1万个字符的输入平均速度快40%，对于超过3.2万个字符的输入成本节省最高可达4倍。 3. 支持高达200万个令牌的Gemini 1.5 Pro也正式发布，支持各种多模态用例。谷歌云提供上下文缓存功能，帮助客户高效利用Gemini 1.5 Pro和Gemini Flash模型的巨大上下文窗口。1. Google Cloud officially releases Gemini 1.5 Flash, offering low latency, competitive pricing, and a groundbreaking 1 million token context window, suitable for a wide range of large-scale use cases from retail chat agents to document processing and research agents. 2. Gemini 1.5 Flash provides significant advantages over comparable models like GPT-3.5 Turbo, with a 60-fold longer context window and 40% faster speed on average for 10,000 character inputs, and up to 4 times cost savings for inputs over 32,000 characters. 3. Gemini 1.5 Pro, supporting up to 2 million tokens, is also officially released, enabling various multimodal use cases. Google Cloud provides context caching functionality to help customers efficiently utilize the vast context windows of Gemini 1.5 Pro and Gemini Flash models.
---
本文由大语言模型（LLM）生成，旨在为读者提供半导体新闻内容的知识扩展（Beta）。
