消息面上，1月29日凌晨，阿里云通义千问旗舰版模型Qwen2.5-Max正式升级发布。据其介绍，Qwen2.5-Max模型是阿里云通义团队对MoE模型的最新探索成果，预训练数据超过20万亿tokens，展现出极强劲的综合性能，在多项公开主流模型评测基 ...

红星资本局注意到，此前1月29日，阿里云公众号发布消息称，阿里云通义千问旗舰版模型Qwen2.5-Max全新升级发布。Qwen2.5-Max模型是阿里云通义团队对MoE模型的最新探索成果，预训练数据超过20万亿tokens。阿里云称，新模型在多项公开主流模型评测基准上录得高分，全面超越了目前全球领先的开源MoE模型以及最大的开源稠密模型。

Explore the impact of DeepSeek's DualPipe Algorithm and Nvidia Corporation's goals in democratizing AI tech for large ...

证券时报e公司讯，科大讯飞在互动平台表示，公司已关注DeepSeek—V3的混合专家（MoE）架构技术进展，公司会持续推动算法迭代，让自主可控的星火大模型在自主可控的算力平台上持续进化提升关系。

ADR shares of Alibaba Group Holding ($BABA) rose more than 6% on Tuesday after the company said its latest artificial ...

Shares of Alibaba (NYSE: NYSE:BABA) climbed 2.8% as the company's latest AI model, Qwen2.5-Max, garnered significant interest ...

Alibaba stock is higher Wednesday after the Chinese conglomerate said its updated AI model outperforms DeepSeek and other ...

Moe believes. Meanwhile, mainland Chinese citizens and entities poured over $100 billion into Hong Kong's stock market last year, showing that "Chinese citizens are buying their own market," he ...