PulseAugur
LIVE 07:50:07
tool · [1 source] · · 中文(ZH) 海光DCU完成与混元Hy3 preview大模型的深度适配
0
tool

Hygon DCU completes deep adaptation with Tencent's new 295B parameter model

Haiguang Information has successfully adapted its DeepComputing 3 DCU with Tencent's latest large model, Hunyuan Hy3 preview. This new Tencent model boasts 295 billion parameters and supports a 256K context window, showing improvements in complex reasoning, agent capabilities, and code generation. AI

Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →

IMPACT Demonstrates hardware-software co-optimization for large models, potentially improving performance and efficiency for AI workloads.

RANK_REASON This is a technical adaptation of an existing model to hardware, not a new model release from a frontier lab. [lever_c_demoted from research: ic=1 ai=1.0]

Read on 36氪 (36Kr) →

COVERAGE [1]

  1. 36氪 (36Kr) TIER_1 中文(ZH) ·

    Huguang DCU Completes Deep Adaptation with Hunyuan Hy3 Preview Large Model

    36氪获悉,近日,海光信息完成深算3号DCU与腾讯混元Hy3 preview大模型的深度适配。据了解,腾讯混元Hy3 preview是腾讯最新发布的大模型版本,具备295B总参数规模,并支持256K超长上下文,在复杂推理、Agent能力和代码生成等方面实现提升。