Skip to main content
ToolPotion

Hy3 Preview

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 performance with a 256K context window. Available on Tencent Cloud and OpenRouter.

View Model
Share

Description

Hy3 Preview, developed by Tencent Hunyuan, is a powerful open-source language model designed to excel in various complex tasks. It features a Mixture-of-Experts (MoE) architecture with 295 billion total parameters, activating 21 billion per inference pass. This design allows it to achieve high performance in math, coding, and multilingual benchmarks, outperforming many larger models.

The model supports a context window of up to 256K, making it suitable for parsing lengthy and complex inputs. Hy3 Preview is available for deployment via vLLM or SGLang and can be accessed through Tencent Cloud and OpenRouter. It offers an OpenAI-compatible API with three reasoning modes: no_think, low, and high.

Hy3 Preview is released under the Tencent Hy Community License Agreement, with model weights available on platforms like Hugging Face, ModelScope, and GitCode. The API access is priced at approximately one-tenth of GPT-4 rates, providing a cost-efficient solution for developers.

The model is integrated into various Tencent products, including Yuanbao, CodeBuddy, and WorkBuddy, enhancing their capabilities with AI-driven features. Hy3 Preview's architecture balances parameter scale and performance, making it a strategic choice for applications requiring efficient and effective AI solutions.

Hy3 Preview Highlights

  • 295B total parameters

  • 21B activated parameters

  • 256K context window

  • Mixture-of-Experts architecture

  • OpenAI-compatible API

  • Available on Tencent Cloud

  • OpenRouter integration

  • Tencent Hy Community License Agreement

  • Competitive benchmark scores

  • Cost-efficient API pricing

  • Integration with Tencent products

  • Supports vLLM and SGLang deployment

  • Three reasoning modes

  • Model weights on Hugging Face

  • Built in 90 days

Getting Started with Hy3 Preview

  1. Configure: Set up deployment via vLLM or SGLang

  2. Use: Access through Tencent Cloud or OpenRouter

  3. Optimise: Select reasoning mode for specific tasks

  4. Integrate: Connect with Tencent products like Yuanbao

Hy3 Preview's Use Cases

  • Math problem solving
  • Code generation
  • Multilingual translation
  • AI chatbot development
  • Productivity tools

FAQ from Hy3 Preview

Popular AI Tools Like Hy3 Preview

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

NVIDIA Nemotron 3 Ultra is a powerful AI model designed for complex reasoning and multilingual tasks. With 550 billion parameters, it excels in long-context analysis and tool use,…

FeaturedAI Models & LLMs

DeepSeek-v3 offers instant AI solutions powered by advanced MoE architecture and state-of-the-art language models. Experience cutting-edge AI capabilities with this stable, free,…

AI Models & LLMs

Command A+ is a Mixture of Experts model with 25B active and 218B total parameters, designed for complex reasoning, vision, and multilingual tasks across 48 languages, providing…

FeaturedAI Models & LLMs

RWKV is a powerful language model that offers efficient inference and flexible fine-tuning capabilities. It supports various applications, including desktop GUIs, web-based…

AI Models & LLMs

AI Models

Olmo from Ai2 is a fully open language model designed for advanced AI research and applications. It offers various model variants optimized for programming, reasoning, and…

FeaturedAI Models & LLMs

AI Models

OpenLLaMA is a permissively licensed, open-source reproduction of Meta AI's LLaMA 7B model. Trained on the RedPajama dataset, it offers 3B, 7B, and 13B parameter versions,…

AI Models & LLMs

Quasar 438B is an OpenAI-compatible LLM inference API offering tool calling, structured output, multi-modality, and speech-to-text capabilities. Developed by Multiverse Computing,…

AI Models & LLMs