Tools3 min read
Microsoft's Bing Team Open-Sources 'Harrier' — a 27B Embedding Model That Tops the Multilingual MTEB Benchmark
Microsoft's Bing team has released Harrier under the MIT license on Hugging Face: a 27-billion-parameter multilingual embedding model that outperforms OpenAI and Amazon on the MTEB v2 benchmark while supporting over 100 languages and a 32,000-token context window. Two smaller variants (0.6B and 270M parameters) make the technology accessible for resource-constrained deployments.