Tencent

• AuK Gen 4 Audio

  Released 4d ago

  A 1.5 billion parameter open source speech foundation model that generates and edits speech through natural language instructions, covering zero shot text to speech, content and acoustic editing, paralinguistic editing,…

  • 🔊 Text to speech
  • 🔊 Audio enhancement
  • 🎤 Voice changing
  • 🗣️ Voice cloning
  • 🔊 Voice enhancement
  • 🎤 Vocal isolation New
• Hy4 preview GGUF Gen 7 Text

  Released 15d ago

  Hunyuan Hy4 preview GGUF is Tencent's quantized release of the Hy4 preview Mixture-of-Experts language model, packaged as two llama.cpp compatible files. The Q4_K_M build is a standard 4 bit quantization at 4.86 bits pe…

  • 💬 Chatting
  • 💻 Coding
  • 🔢 Math New
• Hy4 Gen 7 preview Text

  Released 16d ago

  Preview is a Mixture of Experts language model from Tencent with 770B total parameters, 49B activated per token, and a 1M token context window. It defaults to deep chain of thought reasoning and supports tool calling an…

  • 💬 Chatting
  • 📊 Data analysis
  • 💻 Coding
  • 🤖 Agents
  • 🔬 Scientific research New
• EVIE-Preview-4.5B Gen 3 Multimodal

  Released 27d ago

  Qwen EVIE-Preview-4.5B is a multilingual visual document retrieval model built on the Qwen3.5-4B vision language backbone. It uses ColBERT style late interaction with 128 dimensional multi vector token embeddings to mat…

  • 🔍 Document search
  • 🔍 Information retrieval
  • 🔍 Image search New
• Hy ASR 3.0 Gen 3 Preview Multimodal

  Released 1mo ago

  Preview is Tencent Hunyuan’s speech recognition model for context-aware transcription across Mandarin, English, Cantonese, dialects, noisy audio, and long-form speech.

  • 🗒 Transcription
  • 🔊 Audio transcriptions
  • 🎤 Voice transcription New
• Hy3 295B Gen 7 Text

  Released 2mo ago

  Hunyuan Hy3 is a 295B-parameter Mixture-of-Experts language model with 21B active parameters, developed by the Tencent Hy team. It offers strong reasoning, coding, and agentic capabilities with a 256K token context wind…

  • 💬 Chatting
  • 💻 Coding
  • 🤖 Agents
  • 🚀 Productivity
  • 🔍 Advanced reasoning New
• Hy3 Gen 7 Text

  Released 2mo ago

  Hunyuan Hy3 is a 295 billion parameter Mixture of Experts language model from Tencent with 21 billion active parameters and a 256K token context window. It supports fast and deep reasoning modes, tool calling, and multi…

  • 💬 Chatting
  • 📝 Content creation
  • 💻 Coding
  • 🤖 Agents
  • 🤔 Logical reasoning New
• Hy-MT2 Gen 7 Text

  Released 3mo ago

  Tencent Hunyuan’s open-source multilingual translation model family for fast-thinking, instruction-following translation across 33 languages.

  • 🌍 Translations
  • 🎥 Video translation
  • 🌍 Subtitle translation
• T2I L2P Gen 3 Multimodal

  Released 4mo ago

  Tencent Youtu Research’s open-source pixel-space diffusion transfer method for turning latent text-to-image models into high-quality pixel-generation models.

  • 📷 Images
  • 📸 Photorealistic images
• Pixal3D Gen 3 Multimodal

  Released 4mo ago

  Tencent ARC Lab and collaborators’ image-to-3D generation model for creating high-fidelity 3D assets from a single image. It explicitly lifts pixel features into 3D through back-projection, enabling direct pixel-to-3D c…

  • 🌍 3D images
  • 🎲 3D objects
  • 🎮 3D characters
• Hy3 Gen 3 preview Multimodal

  Released 4mo ago

  Preview is Tencent’s new open-source flagship AI model, built for real-world business use and stronger agentic performance. SCMP reports it has 295B parameters, is Tencent’s most powerful model so far, and is already be…

  • 💬 Chatting
  • 💻 Coding
  • 🤖 Agents
  • 🎩 Sophisticated reasoning
• HY World 2.0 Gen 3 Multimodal

  Released 4mo ago

  HY-World 2.0 is Tencent Hunyuan’s multimodal 3D world model for reconstructing, generating, and simulating explorable worlds. It accepts text, single-view images, multi-view images, and videos, then outputs persistent 3…

  • 🧊 3D
  • 🌎 World creation
  • 📷 3D scanning
  • 🎥 3D scenes

Repositories

• APEX-SQL-Project is an agentic Text-to-SQL framework that shifts the paradigm from passive schema perception . Central to our…
• contextpilot ContextPilot extends context management with planning, structured memory, and soft context offloading. Its co…
• DVQA DVQA - Deep learning-based Video Quality Assessment 12/17/2019 add pretrained model on PGC videos We recommen…
• elephantbench ElephantBench is a closed-book knowledge probe for evaluating whether a language model remembers long-tail fa…
• llm.hunyuan.T1 混元-T1: 强化学习驱动，业内首个超大规模混合Mamba推理模型正式发布! 强化学习在大语言模型的后训练阶段开创了新的Scaling范式，这一突破正日益受到业界重视。随着OpenAI O系列模型与DeepSeek R…

Demos

• Hunyuan-Large Hunyuan-Large模型体验 · running · License unknown
• Hunyuan-T1 Hunyuan T1模型体验 · running · License unknown
• Hunyuan3D-1 Text-to-3D and Image-to-3D Generation · unavailable · License unknown
• Hunyuan3D-2 Text-to-3D and Image-to-3D Generation · running · License unknown
• Hunyuan3D-2.1 Image-to-3D Generation · running · License unknown
• Hunyuan3D-2mini-Turbo Fast Images-to-3D Generation within 1 Second · running · License unknown
• Hunyuan3D-2mv MultiImages-to-3D Generation · running · License unknown
• HY-Motion-1.0 Text-to-3D and Image-to-3D Generation · running · License unknown

Datasets

• Hy-Embodied-0.5-VLA-Data Dataset · License cc-by-4.0 · robotics · 85,994 downloads · by tencent

Papers

• Laya: A LeJEPA Approach to EEG via Latent Prediction over Reconstruction Electroencephalography (EEG) is a widely used tool for studying brain function, with applications in clinical…

• IPO Announced 07 Jun 2004 · Leads: Goldman Sachs
  $184.69M Raised $184.69M Round $790M Post-money 1 investors Closed
• Series B Announced 01 May 2001 · Leads: MIH Holdings, Naspers
  $32M Raised $32M Round 2 investors Closed
• Series A Announced 01 Apr 2000 · Leads: IDG Capital Partners, PCCW
  $2.20M Raised $2.20M Round $11M Post-money 2 investors Closed