AIBase
Home
AI NEWS
AI Tools
GEO & AEO
MCP
AI Models
EN

AI News

View More

Microsoft Bing Team Open Sources Harrier Multilingual Embedding Model

Microsoft Bing team open sources the word embedding model Harrier, which supports over 100 languages and performs excellently in the MTEB v2 benchmark. The model is trained on 2 billion examples and GPT-5 synthetic data, using a 32,000 token context window, with 2.7 billion parameters, significantly improving the accuracy and flexibility of multilingual tasks.

10.4k 39 minutes ago
Microsoft Bing Team Open Sources Harrier Multilingual Embedding Model

Microsoft Bing Team Open Sources 27B Embedding Model Harrier, Top in Multilingual Benchmark Tests

Microsoft Bing's Harrier, a new open-source word embedding model series, outperforms top proprietary models like OpenAI, Amazon, and Google Gemini in multilingual benchmarks. The 27B flagship supports over 100 languages with a 32,000-token context, aiming to transform search, retrieval, and AI agent foundations.....

10.6k 1 hours ago
Microsoft Bing Team Open Sources 27B Embedding Model Harrier, Top in Multilingual Benchmark Tests
AIBase
Empowering the future, your artificial intelligence solution think tank
English简体中文繁體中文にほんご
FirendLinks:
AI Newsletters AI ToolsMCP ServersAI NewsAIBaseLLM LeaderboardAI Ranking
© 2026AIBase
Business CooperationSite Map