5h ago
Feature Request – Allow Merchants to Select Gemini or Seedream Models via OpenRouter
Hi Customily Team,
Since Customily already integrates with OpenRouter, we would like to suggest adding one additional configuration option:
Allow merchants to choose between Gemini and Seedream image-generation models directly from the OpenRouter configuration in the Customily backend.
Our suggested setup is:
If the merchant selects Gemini, Customily can use the currently supported Gemini image model.
If the merchant selects Seedream, the default model should be Seedream 5.0 Flash.
Based on our recent comparison, the reference data is as follows:
Nano Banana 2 Lite
OpenRouter P50 generation time: approximately 6.3 seconds
Image quality: High
Cost: approximately $0.0336 per image (1K)
Seedream 4.5
OpenRouter P50 generation time: approximately 12.5 seconds
Image quality: High
Cost: approximately $0.040 per image
Seedream 5.0 Flash
OpenRouter P50 generation time: approximately 14.1 seconds
Image quality: High
Cost: approximately $0.018 per image and up
Nano Banana 2
OpenRouter P50 generation time: approximately 16.7 seconds
Image quality: Highest
Cost: approximately $0.067 per image (1K)
Seedream 5.0 Pro
OpenRouter P50 generation time: approximately 80.3 seconds
Image quality: Premium
Cost: approximately $0.045 per image and up
Based on these comparisons, some Seedream models are already relatively close to Gemini image models in terms of generation speed and image quality, while offering significantly lower costs.
In particular, Seedream 5.0 Flash provides good image quality and relatively fast generation speed at a cost starting from approximately $0.018 per image, making it very suitable for merchants using AI image generation at scale.
Therefore, we suggest adding a simple Image Model Selector to the existing OpenRouter integration:
Gemini
Seedream
If the merchant selects Seedream, Seedream 5.0 Flash should be used as the default model.
Pending