Lesson Content

Why model selection matters.

Different generation models produce meaningfully different results – different visual quality levels, different motion characteristics, different responsiveness to specific prompt elements, and different credit costs per generation. Choosing a model based on a specific task's requirements rather than always defaulting to the newest or most capable model can produce better-fit results and more efficient credit use.

Key model evaluation dimensions.

Evaluate models on: visual quality (resolution, detail, realism or style), motion quality (smoothness, consistency, camera motion handling), prompt responsiveness (how closely the output follows the prompt versus going its own direction), temporal consistency (how consistent the visual appears across frames), and credit cost per generation.

Runway's evolving model landscape.

Runway has released a series of models – Gen-1, Gen-2, Gen-3 Alpha, the Gen-4 family, and Gen-4.5 at the time of writing, plus the Aleph video editing model – with capabilities expanding significantly between releases. New models and capabilities are added regularly – the current model landscape at the time of any course material may differ from what is available when you are reading this. Always verify the current model lineup and their current characteristics at runwayml.com. This course focuses on evaluation methodology rather than specific current model specs that may be outdated.

Accessing current model information.

Runway's platform provides model-specific documentation, and the creator community produces comparative generation examples. Both are valuable for understanding current model characteristics – direct testing with your specific content type is ultimately the most informative approach.