Unlimited Parallel API Nodes & New Model Options
Recent Model Updates: Minimax Hailuo, Vidu, GPT-5, Kling2.1, Veo3, etc.
We are excited to share a big update for ComfyUI API Nodes! You can now run unlimited API nodes in parallel! Instead of waiting for each request to finish before starting the next, you can trigger multiple API calls at once. It’s the fastest way to test and verify your creative idea!
We’ve also been working on a series of new APIs. Check out the latest available models:
New Models and Example Workflows
MiniMax Hailuo-02 Video
768p for 6-second or 10-second clips.
1080p for crisp 6-second outputs.
Vidu Video Generation
Text-to-Video — bring prompts to life as moving scenes.
Image-to-Video — animate stills with dynamic motion.
Reference-to-Video — guide generation with reference images.
Start-End to Video — define beginning and ending frames, then let the model fill in the motion in between.
Veo 3
veo3 (with/without audio)veo3-fast (with/without audio)
OpenAI Chat (New GPT-5 Family)
gpt-5— full-power reasoning and generation.gpt-5-mini— faster responses with lighter compute.gpt-5-nano— ultra-lightweight for quick interactions and efficient tasks.
Google Gemini (2.5 Series)
gemini-2.5-pro— advanced reasoning and generation.gemini-2.5-flash— optimized for speed and rapid iterations.
Get Started
Update ComfyUI or ComfyUI Desktop to the latest.
Find Minimax Hailuo, Vidu, Gemini, or OpenAI Chat (GPT-5) nodes from Node Library or Templates in the left side bar.
Experiment and compare with parallel generations.
Chain together more complex pipelines without worrying about bottlenecks.
Advanced Example: From Idea to Video in Seconds
Using OpenAI Image Generation for Subject & Background → Kontext Stylization → Gemini Node for Prompt Engineering → Vidu Reference-to-Video Node for video generation.
For more information, check our documentation: https://docs.comfy.org/tutorials/api-nodes/overview
Pricing page: https://docs.comfy.org/tutorials/api-nodes/pricing
As always, enjoy creating!
Is it possible to create custom nodes with parallelism, or is this currently limited to ComfyUI API Nodes? Like creating/forking a custom node that I can use to call other APIs in parallel.