Glossary
CFG scale
Also known as: guidance scale, classifier-free guidance
A parameter controlling how closely the model follows the text prompt. Higher CFG = more literal interpretation; lower CFG = more creative freedom. Typical range 4-12.
In depth
What it means in practice
CFG (Classifier-Free Guidance) scale was originally a Stable Diffusion parameter and is now a standard control across most image models. At low CFG (3-5) the model has freedom to interpret loosely, often producing more aesthetically interesting but less literal output. At high CFG (10-15) the model adheres tightly to every prompt detail, producing literal but sometimes less polished results.
Default values vary by model — typically 7 for SDXL, 4-6 for FLUX, and not exposed for Sora 2 or Kling (which manage guidance internally). For most prosumer use the default is fine; advanced users tune CFG to balance literal adherence against aesthetic quality.
Related
Prompt
The text instruction given to an AI model that describes what to generate. For image and video models, prompts mix subject description, style direction, camera or composition hints, and optional negative prompts.
Seed
A numerical value that initializes the random-number generator inside a model. Fixing the seed makes outputs reproducible — the same prompt + same seed yields the same output.
Inference steps
The number of iterative refinement passes the model takes to produce an output. More steps generally means higher quality, with diminishing returns above a model-specific threshold.