Tencent AI Lab
Tencent AI Lab
VideoOpen Weights (Apache 2.0) Verified Architecture & SpecsFree ($0 API Tokens)

HunyuanVideo Pro

Open-source foundation video generation model featuring a dual-stream Diffusion Transformer (DiT) architecture capable of generating cinematic 1080p 60fps video clips with physically accurate lighting and complex camera motion.

Technical Architecture & Execution Specifications
Architecture Overview

HunyuanVideo Pro

Open-source foundation video generation model featuring a dual-stream Diffusion Transformer (DiT) architecture capable of generating cinematic 1080p 60fps video clips with physically accurate lighting and complex camera motion.

Memory Math Breakdown
  • • FP16 Weights = 13.0B × 2B = 26.00 GB
  • • INT4 Weights = 13.0B × 0.55B = 7.15 GB
  • • KV Cache (16384 ctx, FP16) ≈ 2.00 GB
  • • Activation Buffer = ~20% overhead
Supported Modalities
textimagevideo

Hardware & Execution ParametersVideo

Total Parameter Count13B
Active Parameters (MoE)13B per token
Context Window Capacity16,384 tokens
Model Weights Footprint26.0 GB (FP16) / 7.2 GB (INT4)
Distribution LicenseOpen Weights (Apache 2.0)
Standard API Pricing (1M Tokens)$0 in / $0 out
Model Heritage & Evolutionary Lineage

Genealogical Graph & Evolutionary Provenance

Tracing foundational base architecture ancestry, architectural successors, scale siblings, and reasoning distillation derivatives.

Independent Foundation Checkpoint

HunyuanVideo Pro operates as an autonomous foundation model architecture without direct precursor derivatives in this catalog.

Root Architecture Node
AI Model Architecture & Intelligence

Architecture Engineering & Capability Deep-Dive

An objective architectural evaluation of HunyuanVideo Pro by Tencent AI Lab, analyzing underlying compute dynamics, memory constraints, and deployment economics.

Topology & Attention Mechanics

A robust autoregressive transformer utilizing standard attention patterns for predictable and coherent token generation.

Architecture Type:Advanced Transformer

Evaluation Profile & Reasoning

Exhibits frontier-tier behavior in reasoning and coding.

Domain Specialty:Video

LLM Hardware Sizing & Serving

For open deployments via vLLM/SGLang, quantization (INT4/AWQ) is heavily recommended to fit dense memory constraints, or multi-GPU pipeline parallelism for full FP16.

KV Cache Mgmt: PagedAttention / FlashAttention-3
Hosting Type:Open Weights (Apache 2.0)

Inference Economics & Workflows

Well-suited for enterprise pipelines where capability is balanced against per-million token costs.

Enterprise Fit:Production Ready
Production Trade-Offs & Capability Balance

Architectural Strengths vs. Considerations

An objective balance sheet analyzing the operational advantages and production constraints of deploying HunyuanVideo Pro.

Key Architectural Strengths

  • Ultra cost-effective inference at $0/1M input tokens enables high-frequency agent loops.
  • Demonstrated Dynamic Motion Score evaluation score of 88.1% in verified benchmarks.

Operational Considerations

  • Non-deterministic reasoning chains require schema validation in safety-critical deployments.
LLM Hardware Sizing & Runtime Compatibility

Inference Runtimes & Hardware Sizing

Deployment targets, inference engines, and memory requirements for HunyuanVideo Pro.

Recommended Hardware Profile:
1x RTX 4070 / 4080 (16GB) or Mac M-Series (16GB Unified)
Consumer GPU (< 16 GB VRAM)
Est. 31.2 GB (FP16) / 8.6 GB (INT4)
Production Serving Recipes
vLLM Production
python3 -m vllm.entrypoints.openai.api_server --model HunyuanVideo Pro --tensor-parallel-size 1 --gpu-memory-utilization 0.95 --max-model-len 4096 --enable-chunked-prefill
Ollama / llama.cpp
ollama run hunyuanvideo pro
SGLang Structured
python3 -m sglang.launch_server --model-path HunyuanVideo Pro --tp 1 --trust-remote-code
TGI Serving
text-generation-launcher --model-id HunyuanVideo Pro --num-shard 1 --max-batch-prefill-tokens 32000
Supported Inference Engines
vLLM

High-throughput PagedAttention server

Supported
Ollama

One-click CLI & local desktop serving

Supported
SGLang

Fast multi-turn structured decoding

Supported
TGI

Text Generation Inference

Supported
Llama.cpp

GGUF CPU/Apple Silicon execution

Supported
Precision & Quantization Formats
BF16 / FP16Full Precision

~31.2 GB VRAM required

FP8 (E4M3)Native FP8

~15.6 GB VRAM (Hopper speedup)

AWQ / GPTQ (4-bit)Activation-Aware

~8.6 GB VRAM

GGUF (Q4_K_M)Quantized Binary

CPU RAM / Apple Silicon optimized

LLM Benchmark Database & Performance Metrics

4 Tested

Standardized evaluation results across reasoning, agentic coding, computer use, and alignment.

Flagship Headline MetricsIndustry SOTA Standard
General / Other

Dynamic Motion Score

88.1%
0%100%
General / Other

VBench Video Quality

85.3%
0%100%
General / Other

Text Alignment Score

84.7%
0%100%
General / Other

Physical Consistency Score

81.4%
0%100%
Dynamic Motion Score
88.1%
Text Alignment Score
84.7%
VBench Video Quality
85.3%
Physical Consistency Score
81.4%
Commercial Rates & Inference Costs

API & Deployment Pricing

Open-weights model available for local and private cloud deployment. Compute costs depend on the target GPU hardware instance.

Usage TierRate / Unit
Prompt / Input Tokens$0 / 1M tokens
Completion / Output Tokens$0 / 1M tokens
Inference Cost & ROI Engine
Market Cloud Rate
Prompt / Input Volume:50M Tokens / mo
1M500M1,000M
Generated / Output Volume:10M Tokens / mo
1M250M500M
Estimated Monthly Spend
$13.50/ mo
Input (50M @ $0.15/1M):$7.50
Output (10M @ $0.60/1M):$6.00
Cloud GPU Breakeven Ratio
RunPod RTX 4090 ($316/mo)0.04x spend
Lambda 1x H100 ($1,800/mo)0.01x spend
💡 Open-weights model. You can self-host for $0 token API charge or consume via managed serverless endpoints at the rates shown above.
Similar Frontier Models & Alternatives
Explore All Comparisons

Comparable Foundation Architectures

Alternative models in the Video class with similar capabilities, context windows, or deployment profiles.

Alibaba CloudVideo

Qwen3.8 Flash Next

Context:262k ctx
Parameters:Open Weights
Input Rate:Free / Self-Host
Zhipu AIVideo

GLM-5.3-Flash

Context:1000k ctx
Parameters:Proprietary
Input Rate:Free / Self-Host
Alibaba CloudVideo

Qwen3.8 Flash

Context:1000k ctx
Parameters:Proprietary
Input Rate:$0.15/1M
Integration & Deployment
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ.get("TENCENT_AI_LAB_KEY", "EMPTY"),
    base_url="http://localhost:8000/v1"
)

response = client.chat.completions.create(
    model="hunyuan-video-pro-2026",
    messages=[{"role": "user", "content": "Explain quantum superposition in 2 sentences."}]
)
print(response.choices[0].message.content)
Frequently Asked Questions

Frequently Asked Questions about HunyuanVideo Pro

Essential facts, architectural specs, hardware constraints, and pricing answers for HunyuanVideo Pro.

To run HunyuanVideo Pro (13B) locally, you generally need Depends on quantization. We recommend using quantized GGUF/AWQ formats with Ollama or vLLM to optimize memory footprint.

Primary Sources & Access Repositories

All technical specifications, parameter distributions, context architectures, and benchmark evaluations for HunyuanVideo Pro are audited against primary source release documentation, research whitepapers, and verified vendor API endpoints.