Customily Product PersonalizerCCustomily Product Personalizer
23h ago

Feature Request – Support More AI Image Models and Allow Merchant-Selectable OpenRouter Models

Hi Team, As AI image-generation features become increasingly common, the capability gap between leading image models is becoming smaller. At the same time, as image-generation volume increases, API and token costs are becoming a significant operating expense. We would therefore like to suggest adding support for the following image-generation models: 🥉 Seedream 5.0 Pro ⭐ Qwen Image 3 Pro Both models offer strong image-generation capabilities while also providing significant cost advantages. Based on current approximate pricing: GPT Image 2.5: approximately $0.21 per 1K image Nano Banana 2: approximately $0.067 per 1K image 🥉 Seedream 5.0 Pro: approximately $0.045 per 1K image ⭐ Qwen Image 3 Pro: approximately $0.04 per 1K image 🥉 Seedream 5.0 Pro Seedream 5.0 Pro offers a strong balance between image quality and cost. At approximately $0.045 per image: It is about 79% less expensive than GPT Image 2.5. It is about 33% less expensive than Nano Banana 2. For example, generating 100,000 images would result in an estimated generation cost of approximately $4,500. For POD, personalized products, and other use cases involving large volumes of customized image generation, this cost difference can become very significant at scale. ⭐ Qwen Image 3 Pro Qwen Image 3 Pro offers an even stronger cost advantage. At approximately $0.04 per image: It is about 81% less expensive than GPT Image 2.5. It is about 40% less expensive than Nano Banana 2. For example, generating 100,000 images would result in an estimated generation cost of approximately $4,000. When image-generation volume reaches hundreds of thousands or even millions of images, the cost difference between models becomes increasingly significant. Therefore, allowing users to select their preferred image model would be highly valuable. In addition, if AI models are integrated through a unified model platform such as OpenRouter, it would be very helpful to allow merchants to select the specific image-generation model they want to use instead of being limited to a single default model. This would allow merchants to balance: Image quality Generation speed API cost Specific personalization use cases Different businesses have different priorities. Some scenarios may require the highest possible image quality, while high-volume POD or batch-generation scenarios may prioritize generation speed and cost efficiency. Allowing merchants to select their preferred image model would therefore make AI image-generation functionality more flexible and much better suited for large-scale commercial use. Thank you for your support. We hope this feature request can be evaluated.
PendingPending