In an announcement posted on WeChat, the e-commerce big’s cloud computing and AI arm Alibaba Cloud stated its new Qwen 2.5-Max mannequin additionally outperformed OpenAI’s GPT-4o and Meta Platforms’ Llama-3.1-405B in LLM efficiency benchmark platforms Enviornment-Onerous and LiveBench. Alibaba owns the South China Morning Put up.
The benchmark efficiency of Qwen 2.5-Max, a part of Alibaba’s Tongyi Qianwen LLM household, was on par with Anthropic’s Claude-3.5-Sonnet mannequin, based on Alibaba Cloud. LLMs are the know-how underpinning generative AI providers like ChatGPT.
Alibaba’s multimodal mannequin is obtainable in numerous sizes, from 3 billion to 72 billion parameters, and consists of each base and instruction-tuned variations. The flagship mannequin, Qwen2.5-VL-72B-Instruct, is now accessible by way of the Qwen Chat platform, whereas all the Qwen2.5-VL collection is accessible on open-source platform Hugging Face and Alibaba’s personal open-source group Mannequin Scope.
Parameter is a machine-learning time period for variables current in an AI system throughout coaching, which helps set up how information prompts yield the specified output. Open supply offers public entry to a software program program’s supply code, permitting third-party builders to switch or share its design, repair damaged hyperlinks or scale up its capabilities.