GPT Image 2

OpenAI's GPT Image 2 on fal.ai with explicit quality and output-size controls for one-at-a-time image generation.

Model Overview

GPT Image 2 generates polished images from text prompts. This Nodespell node intentionally exposes a curated subset of controls so pricing stays understandable before the run and can still be corrected to fal's billed amount after completion.

Best At

• Prompt-driven image generation with stronger quality control than a minimal text-to-image surface.
• Clean iteration across fal's published output sizes or explicit custom width and height values.
• Producing exports in PNG, JPEG, or WEBP without extra post-processing.

Limitations / Not Good At

• This node fixes generation to one output image per run.
• Custom sizes still need to satisfy fal's GPT Image 2 limits: multiples of 16, max edge 3840px, aspect ratio up to 3:1, and total pixels between 655,360 and 8,294,400.
• Exact final billing can still vary, so the displayed estimate is only the starting point before post-run correction.

Ideal Use Cases

• Marketing stills, social visuals, and concept art where explicit quality settings matter.
• One-off image generation where predictable starting estimates are more useful than exposing every raw API field.
• Workflows that need a simple OpenAI image model on fal.ai without custom request wiring.

Input & Output Format

• Input: required prompt, plus quality, image_size_preset, optional custom_width / custom_height, and output_format.
• Output: generated image asset returned on response.

Performance Notes

• Higher quality and larger output sizes materially increase cost.
• Custom sizes use the nearest published GPT Image 2 pricing bucket for the starting estimate, then Nodespell corrects credits after the run using fal billing events.