腾讯混元大模型

An official Tencent Hunyuan project entry. This source covers Hunyuan-Large MoE models, pretrained and instruction variants, and inference and fine-tuning guidance.

Features described on the official site

• The model list includes pretrained, instruction and FP8-quantized variants.
• Provides technical reports, model-download links and deployment guidance.
• Links to Hunyuan and Tencent Cloud; model and cloud-service scopes need separate review.

Suggested starting workflow

• Select a specific Hunyuan-Large variant and inference backend.
• Read license, dependency, memory and deployment guidance before running examples.
• Plan evaluation with your own samples and account separately for deployment and cloud API costs.

These are editorial starting suggestions based on product purpose, not workflows tested by this site.

Before you use it

• Not tested hands-on here; feature descriptions come from official material.
• Pricing, regions, languages, commercial use, watermarks and export conditions have not been verified for a single plan.
• The source covers Hunyuan-Large, not the latest full Hunyuan family or a Tencent Cloud API plan. The README describes release of its TRT-LLM backend as a future plan; throughput, memory and benchmark claims are untested vendor statements.