Open Source

Plot twist: the best open-source video model in 2026 came from Tencent. Hunyuan Video 13B dropped with full weights, training code, and a permissive license. The Sora moat just got smaller.

The Setup

Hunyuan Video is a 13B DiT trained jointly on image and video data. The motion is smooth, the prompt adherence is wild, and it slots straight into ComfyUI. There's an official diffusers pipeline if you'd rather script it.

{
  "nodes": {
    "1": { "class_type": "HunyuanVideoLoader", "inputs": { "model": "hunyuan-video-13b" } },
    "2": { "class_type": "CLIPTextEncode", "inputs": { "text": "drone shot over Brisbane river at dawn, cinematic" } },
    "3": { "class_type": "HunyuanVideoSampler", "inputs": {
      "model": ["1", 0], "positive": ["2", 0],
      "frames": 121, "width": 1280, "height": 720, "steps": 30
    }},
    "4": { "class_type": "VideoCombine", "inputs": { "frames": ["3", 0], "fps": 24 } }
  }
}

The Money Pattern

The diffusers route is what I'm using for FFmpeg-driven batch jobs — generate, pipe straight into ffmpeg for format conversion, drop into After Effects for the final colour pass. No Sora credits, no rate limits.

import torch
from diffusers import HunyuanVideoPipeline
from diffusers.utils import export_to_video

pipe = HunyuanVideoPipeline.from_pretrained(
    "tencent/HunyuanVideo",
    torch_dtype=torch.bfloat16,
).to("cuda")

frames = pipe(
    prompt="slow zoom on a brutalist concrete facade, golden hour, 35mm",
    num_frames=121,
    height=720, width=1280,
    num_inference_steps=30,
).frames[0]

export_to_video(frames, "out.mp4", fps=24)

The Catch

5-second clips max at the good quality tier. And 80GB VRAM at full bf16 — that's H100 territory. The fp8 quants squeeze it onto a 4090 but motion gets jittery on complex scenes. Render times are also not joke: 5 minutes for 5 seconds on an H100.

The Verdict

For the first time, open-source video gen is actually usable for short loops, intros, and motion plates. If you're doing After Effects work, this is your new b-roll generator. The 5-second limit is annoying but stitchable. Try it before the inevitable licensing rug-pull discourse starts.