Service category | Service description |
Text vectorization | OpenSearch text embedding service-001 (ops-text-embedding-001): Offers text embedding in over 40 languages, accepting up to 300 characters and producing a 1536-dimensional output vector. OpenSearch general text embedding service-002 (ops-text-embedding-002): Provides text embedding in over 100 languages, with a maximum input length of 8192 characters and a 1024-dimensional output vector. OpenSearch text embedding service-Chinese-001 (ops-text-embedding-zh-001): Specializes in Chinese text embedding, handling up to 1024 characters and delivering a 768-dimensional output vector. OpenSearch text embedding service-English-001 (ops-text-embedding-en-001): Dedicated to English text embedding, it processes up to 512 characters and generates a 768-dimensional output vector.
|
Large model answer services | OpenSearch-Qwen-Turbo (ops-qwen-turbo): Leverages the qwen-turbo large-scale language model, featuring supervised fine-tuning, enhanced retrieval capabilities, and reduced harmful content. Qwen-Turbo (qwen-turbo): An ultra-large-scale language model, Qwen-Turbo supports inputs in various languages, including Chinese and English. For more information, see the Qwen large language model introduction. Qwen-Plus (qwen-plus): An enhanced version of the Qwen ultra-large-scale language model, Qwen-Plus supports multilingual input. For more information, see the Qwen large language model introduction. Qwen-Max (qwen-max): A trillion-level ultra-large-scale language model, Qwen-Max accommodates inputs in various languages. For more information, see the Qwen large language model introduction. Qwen-MAX-LongContext (qwen-max-longcontext): A trillion-level ultra-large-scale language model, this version supports a 30k token context, with API restrictions capping user input at 28k tokens. For more information, see the Qwen large language model introduction.
|