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.

Tencent

AI overview

Sign in and the AI will write an overview from our coverage.

Headlines · 332

  1. LookAtMyQuant 2 Compresses Qwen and Inkling Models

    LookAtMyQuant 2 is an open-source quantization compiler that measures how compression affects a model’s output and assigns codes and bit budgets to stored model components. Its recipe uses post-training quantization, trellis codes, Fisher information for per-row bit allocation, and distillation. The project released a 1.23 GB Qwen3.5-4B artifact and reports compressing Inkling-2-Mini from 18.89 GB of text weights to 5.38 GB and 8.87 GB. The method and MIT-licensed code are available through the project site. The project claims its Gemma 4 E2B GGUF has 12× lower KL divergence than Q4_K_M at the same file size, but the supplied report contains no independent corroboration.

    Hacker News 首页(100+ 分) · 🔥 26
  2. Tencent reportedly pivots toward AI-native agents across its ecosystem

    Tencent may be making a fourth AI strategy shift, moving away from betting on a standalone consumer AI app toward embedding agents across WeChat, mini programs, enterprise services, and other existing products. 木禾商业财经, republished by TMTPost, cites Ma Huateng’s “陆战变海战,陆军变海军” remarks at Tencent’s 28th-anniversary LinkTime meeting as describing a division between traditional businesses and AI-native work. The report says WorkBuddy and WeChat’s Xiaowei would lead the latter, while Hunyuan continues to develop agent, multimodal, and tool-calling capabilities.

    钛媒体 · 🔥 6
  3. Tencent Hunyuan outlines larger and smaller Hy model plans after AMA

    Tencent chief AI scientist Yao Shunyu and the Hunyuan team held a 120-minute Reddit AMA for developers. According to the team, as reported by IT之家, 量子位 and 36氪, Hy4 still has substantial post-training headroom, large-scale reinforcement learning is under way, and larger models are planned. They also said smaller Hy models and a model targeting 24GB of VRAM are on the roadmap.

    IT之家 AI · 🔥 17
  4. Tencent Hunyuan outlines Hy4 roadmap and architecture trade-offs

    Tencent’s Hunyuan team held a two-hour Reddit AMA discussing the Hy4 model family, its architecture, deployment plans, evaluation practices, safety restrictions and product roadmap. The team said the next Hy model will be significantly larger, while Hy4 mini and Flash-Lite were in development. It also acknowledged Hy4’s tendency to overthink and said recursive self-improvement remains at an early stage. The AMA followed Tencent’s opening of Hy4’s weights and release of Hy4 flash; reported cost and ranking figures came from Tencent disclosures or third-party rankings.

    钛媒体 · 🔥 16
  5. Bloomberg: AI-exposed China job postings declined faster

    A Bloomberg News analysis of nearly 10 million Chinese online job postings from July 2021 to June 2025 found that listings for roles highly exposed to AI fell about 10 percentage points more than listings for less affected roles. In nearly every industry, declines were steeper where job tasks overlapped more heavily with AI capabilities, according to a Peking University index. Bloomberg also cited two independent Chinese studies that reached similar conclusions. It cautioned that the analysis cannot isolate AI’s causal effect because China’s broader economic slowdown may have disproportionately reduced office-job postings.

    The Decoder · 🔥 4
  6. Tencent WorkBuddy adds computer control in public beta

    Tencent announced that WorkBuddy has added Computer Use capabilities for recognizing pages and client interfaces, then clicking, typing and scrolling without requiring target software to open APIs. The feature is in public beta for macOS and Windows and requires WorkBuddy client version 5.11.0 or later, activation and authorization. Users can describe a task in the chat window and view progress in a side panel. Tencent advises users to remain present and monitor sensitive actions, including login, authorization, payment, submission and deletion. The company said interface-recognition results may vary by software.

    IT之家 AI · 🔥 4
  7. Chinese tech giants diverge on AI spending and revenue

    A 36Kr analysis of Alibaba, Baidu, Tencent and JD.com’s 2026 third-quarter results found different AI revenue and investment models. Alibaba reported RMB60.973 billion in AI cloud and computing-services revenue, up 53% year on year, while Baidu reported RMB13.2 billion in AI core-business revenue. The companies did not clearly isolate AI’s contribution to advertising, e-commerce or supply-chain growth. Investment also affected results: Alibaba’s AI labs and applications recorded a wider adjusted EBITA loss than a year earlier, while Tencent made RMB65.1 billion in capital-expenditure cash payments, including substantial advance purchases of computing capacity. JD.com reported that AI-related research and development would put short-term pressure on its margins.

    36氪 人工智能 · 🔥 4
  8. Insta360 Mic Pro and Wave gain native WorkBuddy integration

    Insta360 announced that a free firmware upgrade lets Mic Pro and Wave connect natively with Tencent WorkBuddy. Mic Pro recordings can transfer to WorkBuddy for transcription, summaries, cross-meeting analysis and document generation, while Wave can launch cloud tasks and produce meeting notes, action items and to-do lists. Insta360 and Tencent describe the integration as the product-level implementation of their ecosystem partnership.

    品玩 实时要闻 · 🔥 7
  9. Tencent WorkBuddy launches Computer Use public beta

    Tencent WorkBuddy has launched a public beta of its Computer Use capability for macOS and Windows, according to 36Kr. The feature can recognize current pages and client interfaces and perform background actions such as clicking, typing and scrolling without requiring target applications to expose APIs. Users must update the client to version 5.11.0 or later, enable the feature and complete the relevant authorization. Interface-recognition performance may vary by application.

    36氪快讯 · 🔥 4
  10. Yao Shunyu details foundation-model engineering priorities at Tsinghua

    According to 36Kr AI, Tencent chief AI scientist Yao Shunyu held a closed-door exchange at Tsinghua University’s Yao Class on November 18, covering foundation-model infrastructure, training algorithms, data, safety, organization and commercialization. He questioned metric-driven evaluation and “horse-race” resource allocation, arguing that independent evaluations, data-quality judgment, engineering rigor and organizational trust remain central to model development. He also said concepts such as RSI should return to concrete engineering problems, while startups should reassess how their businesses relate to advances in foundation models.

    36氪 人工智能 · 🔥 5
  11. Tencent publishes NVFP4 WeMM-Video-35B-A3B checkpoint

    Tencent published the WeMM-Video-35BA3B-NVFP4 checkpoint, an NVFP4 W4A16 Safetensors variant of WeMM-Video-35B-A3B with 24.11 GiB of weight files. The model card says it supports Chinese and English audio-visual understanding, accepting text, images, audio, and video, and includes instructions for serving it with patched vLLM 0.31.0, ModelOpt Mixed Precision, and Marlin.

    腾讯混元 · Hugging Face · 🔥 3
  12. Tencent releases WeMM-Video-35B-A3B GGUF bundle

    Tencent’s WeMM-Video model repository introduced an Apache-2.0 GGUF Q4_K_M bundle for WeMM-Video-35B-A3B, with separate language, vision, and audio components. The repository says the WeMM native CPU runtime loaded the bundle and completed generation checks for text, images, audio, and video with a soundtrack; Tencent describes the model family as Chinese-English bilingual and capable of joint audio-visual understanding and reasoning.

    腾讯混元 · Hugging Face · 🔥 3
  13. Tencent releases FP8 checkpoint for WeMM-Video-35B-A3B

    Tencent’s Hugging Face model repository lists a new FP8 Block128 Safetensors variant of WeMM-Video-35B-A3B under the ID tencent/WeMM-Video-35BA3B-FP8. The checkpoint is based on the quantized 35B-A3B model, uses Apache-2.0 licensing, and contains 36.09 GiB of weight files. Tencent says the bilingual model accepts text, images, audio, and video, with loading and generation checks reported on an NVIDIA L20 and H20.

    腾讯混元 · Hugging Face · 🔥 3
  14. WeChat Tightens AI Rules for Short Dramas

    DoNews reports that WeChat’s short-drama review rules will raise the maximum episode duration to one minute. The report also says that, starting December 1, AI-generated proof videos will be required for short dramas submitted to the platform.

    DoNews · 🔥 3
  15. VeloxCon China 2026 to Focus on AI Execution Infrastructure

    A report says VeloxCon China 2026 is scheduled for December 5 at Alibaba’s Xuhui Binjiang campus in Shanghai. The reported agenda includes GPU acceleration, agent infrastructure and unified multi-engine architecture. It also lists experts from Meta, Alibaba Cloud, NVIDIA, IBM, Google, Tencent and Ant, though their participation is not independently confirmed in the supplied report.

    掘金 · 人工智能 · 🔥 1
  16. Tencent releases AWQ-quantized WeMM-Video-35B-A3B

    Tencent published an AWQ W4A16 Safetensors checkpoint for WeMM-Video-35B-A3B, with 22.69 GiB of weight files and INT4 feed-forward weights. The bilingual audio-visual model accepts text, images, audio, and video; its documentation includes Transformers loading instructions, patched vLLM 0.18.0 serving guidance, and validation on NVIDIA L20 and H20 hardware.

    腾讯混元 · Hugging Face · 🔥 3
  17. Tencent releases WeMM-Video-9B multimodal checkpoint

    Tencent's model card describes WeMM-Video-9B as the dense 9B checkpoint in the bilingual WeMM-Video family for joint audio-visual understanding and reasoning. It accepts text, images, audio, and video and generates text responses. The release includes BF16 Safetensors, quantized variants, Transformers support, and serving instructions for patched vLLM and SGLang versions.

    腾讯混元 · Hugging Face · 🔥 3
  18. Tencent releases FP8 Block128 checkpoint for WeMM-Video-9B

    Tencent released the FP8 Block128 Safetensors variant of WeMM-Video-9B under the Apache 2.0 license, with 13.10 GiB of weight files available on Hugging Face. The Chinese-English bilingual model accepts text, images, audio, and video and generates text responses; its model card includes Transformers loading instructions and a patched vLLM 0.18.0 serving setup. Tencent reports successful loading and generation checks on NVIDIA L20 and H20 GPUs, but these validation details are self-reported.

    腾讯混元 · Hugging Face · 🔥 3
  19. Tencent releases WeMM-Video-9B Q4_K_M multimodal bundle

    Tencent published a GGUF Q4_K_M bundle for WeMM-Video-9B on its Hugging Face repository. The package includes separate language, vision, and audio components and is intended for the WeMM native runtime, with support for text, image, audio, and video inputs as described by Tencent. The repository also says the CPU runtime completed generation checks across these input formats.

    腾讯混元 · Hugging Face · 🔥 3
  20. Tencent releases NVFP4-quantized WeMM-Video-9B checkpoint

    Tencent published an NVFP4 W4A16 Safetensors variant of WeMM-Video-9B, with 13.07 GiB of weight files. The checkpoint accepts text, images, audio and video and generates Chinese or English responses, according to Tencent's Hugging Face repository; deployment uses patched vLLM 0.31.0 with ModelOpt Mixed Precision and Marlin.

    腾讯混元 · Hugging Face · 🔥 3
  21. Tencent releases WeMM-Video-9B AWQ W4A16 checkpoint

    Tencent’s Hugging Face repository released a 12.86 GiB AWQ W4A16 Safetensors checkpoint for WeMM-Video-9B. Tencent describes it as a Chinese-English bilingual model that jointly processes text, images, audio, and video, with support for Transformers and patched vLLM 0.18.0. The repository reports successful loading and generation checks on NVIDIA L20 and H20 hardware, but these validation details come from the publisher.

    腾讯混元 · Hugging Face · 🔥 3
  22. Tencent releases WeMM-Video-4B NVFP4 checkpoint

    Tencent’s model repository publishes WeMM-Video-4B-NVFP4, an Apache 2.0-licensed NVFP4 W4A16 Safetensors derivative of WeMM-Video-4B with 7.42 GiB of weights. The bilingual video-text-to-text model accepts text, images, audio, and video, and the repository provides patched vLLM 0.31.0 serving instructions using ModelOpt Mixed Precision and Marlin, plus limited generation checks.

    腾讯混元 · Hugging Face · 🔥 3
  23. WeMM-Video-4B model card details multimodal checkpoint

    The Hugging Face model card for tencent/WeMM-Video-4B describes the checkpoint as the dense 4B version in the bilingual WeMM-Video family. It says the model accepts text, images, audio, and video and generates Chinese or English responses. The listing provides BF16 Safetensors weights, four quantized variants, Apache License 2.0 terms, and examples for Transformers, vLLM, and SGLang deployment.

    腾讯混元 · Hugging Face · 🔥 3
  24. Tencent releases WeMM-Video-4B GGUF Q4_K_M bundle

    Tencent’s WeMM-Video-4B-GGUF repository provides an Apache 2.0-licensed Q4_K_M bundle for the WeMM native runtime. It contains separate language, vision, and audio GGUF components sized at 4.90 GiB, 0.63 GiB, and 2.00 GiB, respectively, and the repository says all three must use the same Hub revision. Tencent describes the model as a Chinese-English bilingual system for joint audio-visual understanding that accepts text, images, audio, and video.

    腾讯混元 · Hugging Face · 🔥 3
  25. Tencent releases WeMM-Video-35B-A3B multimodal checkpoint

    Tencent’s WeMM-Video model card says WeMM-Video-35B-A3B is a bilingual Chinese-English mixture-of-experts checkpoint with 35B total parameters and 3B active parameters. It accepts text, images, audio, and video, and the Hugging Face repository publishes BF16 weights plus FP8, AWQ, GGUF, and NVFP4 variants under an Apache 2.0 license. Tencent claims the model leads open-source omni-modal video understanding, but that benchmark claim is vendor-reported.

    腾讯混元 · Hugging Face · 🔥 3
  26. Tsinghua Yao Class alumni form a prominent global AI network

    A 36Kr article republished from Phoenix Weekly uses Tencent chief AI scientist Yao Shunyu’s November 18 lecture at Tsinghua’s Yao Class to examine the school’s role in AI talent. It identifies alumni in AI research and leadership roles at Tencent, DeepSeek, OpenAI, Meta and xAI, as well as in academia, infrastructure and applied technology. The report says claims that Yao Class alumni earn exceptional salaries, including claims about Yao Shunyu’s Tencent compensation and DeepSeek internships, have been debunked.

    36氪 人工智能 · 🔥 3
  27. Tencent Releases FP8 WeMM-Video-4B Checkpoint

    Tencent published an FP8 Block128 Safetensors variant of WeMM-Video-4B under the Apache 2.0 license. The model card documents loading and generation checks on one NVIDIA L20 with Transformers 5.2.0 and PyTorch 2.10.0+cu129, alongside serving instructions for patched vLLM 0.18.0 on FP8-capable GPUs. Tencent describes WeMM-Video as a Chinese-English bilingual family for joint audio-visual understanding and reasoning that accepts text, images, audio, and video.

    腾讯混元 · Hugging Face · 🔥 3
  28. Tencent releases WeMM-Video-4B AWQ W4A16 checkpoint

    Tencent published a 7.32 GiB AWQ W4A16 Safetensors checkpoint for WeMM-Video-4B. The language model’s feed-forward weights use group-128 INT4, while activations and attention weights remain BF16. The repository provides direct-loading instructions for Transformers 5.2.0 and a serving guide for patched vLLM 0.18.0 with the checkpoint’s compressed-tensors configuration; Transformers loading requires compressed-tensors 0.13.0. Tencent reports that the local checkpoint loaded and completed generation checks for text, images, audio, video frames and video with its soundtrack on one NVIDIA L20 using Transformers 5.2.0 and PyTorch 2.10.0+cu129.

    腾讯混元 · Hugging Face · 🔥 3
  29. Tencent releases WeMM-Video-2B GGUF multimodal bundle

    Tencent published a GGUF Q4_K_M bundle for WeMM-Video-2B, providing language, vision, and audio components for the WeMM native runtime. The package supports text, image, audio, video, and combined video-audio prompts under the Apache 2.0 license. Tencent says the CPU runtime loaded the bundle and passed generation checks across those input modes.

    腾讯混元 · Hugging Face · 🔥 3
  30. Tencent releases WeMM-Video-2B-FP8 checkpoint

    Tencent released the WeMM-Video-2B-FP8 repository, an FP8 Block128 Safetensors variant of WeMM-Video-2B with 4.84 GiB of weights. The model supports text, image, audio, and video inputs and generates Chinese or English text responses; its listed capabilities come from Tencent’s model description. The repository documents successful loading and generation with Transformers 5.2.0 and patched vLLM 0.18.0 on an NVIDIA L20.

    腾讯混元 · Hugging Face · 🔥 3

Next →

Experience and discussion from the community

Share my Tencent experienceAsk about Tencent

CubeSandbox: A secure, lightweight sandbox service for AI agents with E2B SDK compatibility.

What problem it solves AI agents need isolated environments that can execute code and workloads while supporting fast creation, secure network access, and manageable state. They also need an interface that can integrate with existing agent tooling and deployment workflows. Key features Hardware-level isolation — Each sandbox runs in its own MicroVM with a dedicated operating system kernel. Kernel-…

github.comhttps://github.com/TencentCloud/CubeSandbox

Hy-MT2: A family of fast-thinking multilingual translation models for complex real-world translation scenarios.

What problem it solves Real-world translation involves more than converting text between languages: tasks can require terminology, style, personalization, delimiter preservation, and structured-data constraints. Hy-MT2 is designed to support translation among 33 languages and follow translation instructions in multiple languages for complex real-world scenarios. Key features Instruction-driven tas…

github.comhttps://github.com/Tencent-Hunyuan/Hy-MT2

LoopForge: LoopForge is a self-hosted workspace where your team ships software with coding agents.

What problem it solves A coding-agent run can skip planning, omit tests, review its own work, or stop before completion, making delivery inconsistent and difficult to supervise. Teams need repeatable workflows that separate roles, execute work in sandboxes, preserve review evidence, and require human approval before changes ship. Key features Repeatable SOP delivery — Standard operating procedures…

github.comhttps://github.com/Tencent/LoopForge

WeKnora: An open-source, LLM-powered knowledge framework for enterprise document understanding, semantic retrieval and reasoning.

What problem it solves Teams need to bring documents together so they can be searched, reasoned over and kept up to date. They also need a way to move from source-grounded retrieval to multi-step tasks while maintaining organized, current knowledge. Key features Three knowledge workflows — RAG supports lookup, an agent handles multi-step tasks, and a wiki organizes knowledge. All three operate on …

github.comhttps://github.com/Tencent/WeKnora