Use Combo Mode to combine two image models

Combo Mode is not mainly a way to repair a messy prompt. It is a model-combination workflow: use one image model for the base shot, then pass that image into a second model chosen for a different strength. For example, Step 1 might give you the best realism and anatomy, while Step 2 can improve lighting, texture, detail, or final polish.

Section prompting: give both models a clean brief

Section prompting still matters because both models need a clear target. A long single-line prompt can bury important details, while labeled sections keep the subject, pose, wardrobe, scene, style, and camera direction easier to preserve through the chain.

• Subject: who is in the frame - your character and their identity.
• Pose: body angle, posture, expression, and what they are doing.
• Wardrobe: clothing, accessories, and outfit details.
• Scene: where it happens - setting, environment, time of day.
• Style: lighting, mood, color, and art direction.
• Camera: lens, framing, shot type, and angle.

Combo Mode: combine model strengths

Combo Mode runs your generation in two passes. The first model creates the source image from your prompt and references. SomaLabs then sends that image to a Step 2 model, so the final output can benefit from one model's base realism or composition and another model's lighting, style, texture, or detail handling.

1. Pick Step 1: choose the model that gives the strongest base for the shot - realism, anatomy, character likeness, or composition.
2. Pass one: the first model generates an image from your section prompt and any references.
3. Pick Step 2: choose a second model for the strength you want to add, such as lighting, texture, color, detail, or a cleaner finish.
4. Pass two: the second model uses the first image plus your character references, aiming to preserve the useful structure while improving the final look.
5. Delivery: the refined image is your final result. If Step 1 succeeds but Step 2 cannot complete, SomaLabs falls back to the Step 1 image and returns the Step 2 cost.

Put it together

Write a clear section prompt, attach your character, choose the Step 1 model for the base image, then choose the Step 2 model for the improvement you want. Combo Mode is most useful when the first model gets the core shot right but a second model can improve the lighting, texture, detail, or polish.

Combo Mode is included on every paid plan. If you are deciding which tier fits your volume, the pricing page lays out monthly credits and what each plan includes.