Lesson Content

When to use the most capable model.

The most capable model (typically the newest, highest quality, most credit-intensive) is appropriate when: output quality is critical and visible, the content requires the best photorealism or temporal consistency available, and the project budget supports higher credit cost per generation. For client deliverables, final production content, or content where visual quality is the key value proposition, the highest-capability model is typically the right choice.

When a lower-capability model is appropriate.

Lower-capability models (faster, less credit-intensive) are appropriate for: concept exploration and ideation (testing prompt directions before committing to expensive final generation), internal assets not requiring maximum quality, content where a stylized or less-photorealistic aesthetic is actually preferred, and batch generation of many variations for selection.

Credit efficiency decision-making.

Running 10 concept explorations at lower model cost versus 3 at highest model cost gives more variation to select from at the same total credit cost. For iterative prompting work where many generations are needed to find the right direction, credit-efficient models maximize the number of iterations possible within a budget.

Documenting model selections.

For projects spanning multiple sessions (or for a team with multiple contributors), document which model was used for each production generation so all clips in a series maintain consistent visual characteristics – mixing models across clips in a sequence can produce noticeable inconsistency in visual style.