Qwen Image 2 Pro (fal.ai)

FAL_QwenImage2Pro is the "Pro" tier of the same API-only Qwen Image 2 model, served from fal's fal-ai/qwen-image-2/pro/text-to-image endpoint. The controls are identical to the standard FAL_QwenImage2 node - same prompt, same image_size presets, same enable_prompt_expansion - so the difference is what fal's Pro tier actually is: the higher-quality, higher-priced lane of the same model family. Think of it as the Qwen answer to Nano Banana Pro: you pay more per call for the more demanding version of the same architecture.

It exists for the same reason every Pro tier exists. If your results from the base node are good enough, don't spend the extra money. If the base tier's output is close but not landing - or you need the premium lane's consistency for a batch - this is the node.

How it works

Same mechanism as every node in this pack: prompt goes to fal, fal.subscribe waits on the endpoint, result images are downloaded and returned as a batched IMAGE tensor. Nothing runs locally, nothing downloads.

The inputs that matter:

• prompt (required) - with no diffusion dials on this architecture, this is the entire game.
• image_size - presets (square_hd default) or custom, with custom_width/custom_height limited to 512–2048px (the model's native 2K design).
• enable_prompt_expansion (default true) - internal prompt rewriting; the main quality lever here. Turn it off when you want your exact wording respected.
• enable_safety_checker (default on), negative_prompt (sent only when non-empty), num_images (1–4), seed (-1 random), output_format (png/jpeg/webp), private_mode.

One IMAGE output into a Preview, Save, or the next stage of your workflow.

Install

cd ComfyUI/custom_nodes
git clone https://github.com/elle-trudgett/fal-ai-comfyui-nodes
pip install -r fal-ai-comfyui-nodes/requirements.txt

or ComfyUI Manager → ComfyUI fal.ai Nodes. No model files - the weights don't exist for download. Set your fal key: export FAL_KEY="...", or a config.ini with [API] / FAL_KEY in the node folder for ComfyUI Desktop; get keys at fal.ai/dashboard/keys and restart.

Gotchas

• "FAL_KEY not set" - env var or config.ini, restart.
• The Pro tax is real. Pro-tier generations bill more per call, and 2K-sized output bills more again. Prototype on the base node, then promote.
• No steps or guidance knobs. This architecture doesn't expose diffusion-style settings; prompt and expansion are your levers.
• Errors arrive as fal.ai error: ... with the API's message intact.

Honest advice: most people should try the standard FAL_QwenImage2 first and only reach for the Pro lane when they can point at a specific quality gap. If you can't tell the difference in a side-by-side, your money is better spent on more iterations of the base tier.