Skip to main content
ToolPotion
AI Advisor
Submit a Tool (Free)

Hy3

LLMs#24AI Models & LLMsOpen source · Web

Hy3 is a 295B-parameter Mixture-of-Experts model developed by the Tencent Hy Team. It excels in reasoning, agentic, and long-context tasks, offering reliable performance across various productivity scenarios.

Visit Website
Share
Hy3 screenshot
Hy3 screenshot

Hy3 is a sophisticated Mixture-of-Experts (MoE) model developed by the Tencent Hy Team, featuring 295 billion total parameters with 21 billion active parameters and 3.8 billion MTP layer parameters. Launched after the Hy3 Preview, the model has been refined with feedback from over 50 products and enhanced post-training with high-quality data. Hy3 outperforms similar-sized models and rivals flagship open-source models with 2-5 times more parameters.

Read More

The model demonstrates significant utility gains across various productivity tasks, including coding, office work, financial modeling, frontend design, and game development. It is a cost-effective option for those seeking reliable performance in these areas. Hy3's architecture includes 80 layers, one MTP layer, 64 attention heads, and a hidden size of 4096, supporting a context length of 256K and a vocabulary size of 120,832.

Hy3's performance is not fully captured by public benchmarks, so a blind evaluation with 270 experts was conducted, where it scored 2.67 out of 4, outperforming GLM-5.1. The model excels in frontend development, data & storage, and CI/CD tasks. Improvements in stability, knowledge retention, and multi-turn intent tracking have been made, reducing hallucination rates and commonsense errors significantly.

Hy3 is open-sourced, with model weights available on platforms like Hugging Face, ModelScope, GitCode, and CNB. It supports deployment with vLLM or SGLang and offers a comprehensive model finetuning pipeline. The model is released under the Apache License 2.0.

  • CategoryAI Models & LLMs
  • Pricing ModelOpen source
  • DeploymentCloud
  • PlatformsWeb

  • 295B total parameters

  • Mixture-of-Experts architecture

  • 21B active parameters

  • 3.8B MTP layer parameters

  • 80 layers excluding MTP

  • 64 attention heads

  • Context length of 256K

  • Vocabulary size of 120,832

  • Open source

  • Fine-tuning support

  1. Access model page: Visit the Hy3 page on Hugging Face

  2. Authenticate: Log in to your Hugging Face account

  3. Configure: Set model parameters and preferences

  4. Send prompts: Interact with the model using your input

  5. Fine-tune: Adjust the model for specific tasks

  • Coding Assistance
  • Office Automation
  • Financial Modeling
  • Frontend Design
  • Game Development

Popular AI Tools Like Hy3

Hy4 Preview is a new-generation Mixture-of-Experts model developed by the Tencent Hy Team. It features 770 billion parameters and is designed for productivity across various…

AI Models & LLMs

Ranked #34 in LLMs

View Tool

New-generation Mixture-of-Experts model developed by the Tencent Hy Team. It features 770 billion parameters and is designed for productivity across various…Hy4 Preview is a new-generation Mixture-of-Experts model developed by the Tencent Hy Team. It features 770 billion parameters and is designed for productivity across various…

AI Models & LLMs
Ranked #34 in LLMs

Hy3 Preview

AI Models

Hy3 Preview by Tencent Hunyuan is an open-source language model featuring 295 billion parameters. It excels in math, coding, and multilingual tasks, offering competitive…

AI Models & LLMs
View Tool

AI Models

Hy3 Preview by Tencent Hunyuan is an open-source language model featuring 295 billion parameters. It excels in math, coding, and multilingual tasks, offering competitive…

AI Models & LLMs
See the Top 5 AI Models

Kimi-K2-Instruct is a state-of-the-art mixture-of-experts language model designed for advanced AI tasks. It excels in reasoning, coding, and tool use, offering robust performance…

AI Models & LLMs

Ranked #44 in LLMs

View Tool

State-of-the-art mixture-of-experts language model designed for advanced AI tasks. It excels in reasoning, coding, and tool use, offering robust performance…Kimi-K2-Instruct is a state-of-the-art mixture-of-experts language model designed for advanced AI tasks. It excels in reasoning, coding, and tool use, offering robust performance…

AI Models & LLMs
Ranked #44 in LLMs

Ling-3.0-tiny is a lightweight hybrid reasoning model with 7.9B parameters, designed for efficient local deployment. It supports fast responses and multi-step reasoning, making it…

AI Models & LLMs
View Tool

Lightweight hybrid reasoning model with 7.9B parameters, designed for efficient local deployment. It supports fast responses and multi-step reasoning, making it…Ling-3.0-tiny is a lightweight hybrid reasoning model with 7.9B parameters, designed for efficient local deployment. It supports fast responses and multi-step reasoning, making it…

AI Models & LLMs
See the Top 50 LLMs

Kimi K2.7 Code is a coding-focused agentic model that enhances software engineering workflows. It improves token efficiency and task completion, offering substantial advancements…

AI Coding Assistants

Ranked #17 in LLMs

View Tool

Coding-focused agentic model that enhances software engineering workflows. It improves token efficiency and task completion, offering substantial advancements…Kimi K2.7 Code is a coding-focused agentic model that enhances software engineering workflows. It improves token efficiency and task completion, offering substantial advancements…

AI Coding Assistants
Ranked #17 in LLMs

Laguna M.1 is a 225B parameter Mixture-of-Experts model designed for agentic coding and long-horizon tasks. It features high-capacity expert routing and global attention…

AI Models & LLMs

Ranked #35 in LLMs

View Tool

225B parameter Mixture-of-Experts model designed for agentic coding and long-horizon tasks. It features high-capacity expert routing and global attention…Laguna M.1 is a 225B parameter Mixture-of-Experts model designed for agentic coding and long-horizon tasks. It features high-capacity expert routing and global attention…

AI Models & LLMs
Ranked #35 in LLMs

Mistral Large 3 is a state-of-the-art AI model designed for enterprises, enabling customization, fine-tuning, and deployment of AI assistants and agents. It features a sparse…

AI Models & LLMs
View Tool

State-of-the-art AI model designed for enterprises, enabling customization, fine-tuning, and deployment of AI assistants and agents. It features a sparse…Mistral Large 3 is a state-of-the-art AI model designed for enterprises, enabling customization, fine-tuning, and deployment of AI assistants and agents. It features a sparse…

FeaturedAI Models & LLMs
See the Top 5 AI Models

Ling-3.0-flash is a next-generation hybrid reasoning model designed for efficiency and performance. With 124 billion parameters, it excels in reasoning, instruction following, and…

AI Models & LLMs
View Tool

Next-generation hybrid reasoning model designed for efficiency and performance. With 124 billion parameters, it excels in reasoning, instruction following, and…Ling-3.0-flash is a next-generation hybrid reasoning model designed for efficiency and performance. With 124 billion parameters, it excels in reasoning, instruction following, and…

AI Models & LLMs
See the Top 50 LLMs
1 of 8