Tencent Hy3 Preview 295B

Overview

Tencent's frontier preview: 295B MoE with 21B active params plus a 3.8B MTP module, 80 layers, top-8 of 192 experts, with fused fast/slow thinking. Released April 2026 under the custom Hunyuan license.

When to pick this model

• Research on fused fast/slow-thinking architectures
• Long-context workloads up to 256k tokens
• Base or Instruct fine-tuning at frontier scale
• Deployments where Tencent's Hunyuan license is acceptable
• Comparing Chinese hyperscaler frontier weights

VRAM requirements by quantization

VRAM figures include model weights plus a typical 8k KV cache and ~600 MB runtime overhead (Ollama / llama.cpp baseline). Add headroom for higher context lengths.

Strengths

• Tencent's first frontier-scale open-weight release
• 256k context window
• Both Base and Instruct variants shipped
• MTP module accelerates long-form generation
• Fused fast/slow thinking in one model

Limitations

• Custom Tencent Hunyuan Community License — legal review required
• Around 177 GB VRAM in Q4
• No Ollama support at launch
• Preview status means rough edges in tooling

Architecture & training

Architecture: MoE 295B/21B active · 80 layers + 1 MTP layer · 192 experts top-8 · GQA 64Q/8KV · BF16

Training: Fused fast/slow-thinking.

Quick start

# HuggingFace : tencent/Hy3-preview

Or use the open-source MCP server to query this model from Claude Desktop, Cursor, or any MCP-compatible client.