Create

Sign in to ReadmeX

or

How we handle your data: Privacy Policy

ReadmeX

Your AI reading companion

What would you like to catch up on?

Pick a briefing, or ask your own question.

Your community and people briefings stay personal to you.

Story

Tencent launches Hy4 flash for lower-cost agent workloads

AI summary

Tencent has launched Hy4 flash, a model with 224B total parameters, 12B active parameters, multimodal support and a 1M-token context window. APPSO reported that it completed WorkBuddy tasks faster and at lower cost than DeepSeek V4.1 Flash, but sometimes produced fewer details; Tencent prices input at RMB 0.5 and output at RMB 2 per million tokens.

In-depth insight

Hy4 flash positions Tencent’s agent and productivity capabilities for frequent, cost-sensitive use, while APPSO’s comparison highlights the trade-off between faster delivery and sometimes less detailed output.

11 月初,腾讯 Hy4 发布,新模型在大模型竞技场上一跃升至开源第一。

排在 Hy4 前面的三个模型分别是 GPT-6 Sol(Max)、Claude Opus 5(Max)和 Fable 5(High),附近还有 GPT-6.1 Terra、GPT-5.6 Sol 等一众强劲的闭源模型。

Image

我们此前在 WorkBuddy 内体验过 Hy4 的 Agent 长程任务和日常办公能力。从逐帧分析视频、复刻游戏和网页,到文件整理、Office 文档处理,Hy4 都交出了不错的成果。

整体实测表现,也基本符合它在大模型竞技场上的成绩。

Some source content is omitted. Continue reading on the original site.

#Tencent#Hy4 flash#Hy4#WorkBuddy#DeepSeek V4.1 Flash

Original · 9 min readRead the original
Group chat

No comments yet. Start the conversation.