AI Models Guide
Understand the differences between AI models and choose the right one for your store.
AI Product Tools supports a wide range of AI models from different providers. Each model has its own strengths, weaknesses, and costs. Choosing the right model depends on whether you prioritize speed, quality, or cost-efficiency.
Overview of AI Models
Different AI models are designed for different tasks. Some are "heavy" and highly intelligent, while others are "light" and extremely fast.
OpenAI Models
OpenAI offers some of the most versatile and widely used models in the world.
- Default in AI Product Tools:
gpt-4.1-mini, a balanced everyday model for product content. - Higher-capability GPT models: Useful for important descriptions, complex prompts, or technical products when your OpenAI account has access.
- Reasoning-focused models: Useful for complex product data or structured decision-making, but may be slower or more expensive.
Best for: Versatile content, wide language support, and high-end creative writing.
Google Gemini Models
Google's Gemini models are known for their incredible speed and large context windows.
- Default in AI Product Tools:
gemini-2.5-flash, a fast model for testing and high-volume work. - Gemini Pro models: Good for deeper product context when your Google account has access.
- Gemini Flash models: Best when speed and cost-efficiency are the priority.
Best for: Speed, cost efficiency, and processing large amounts of product data quickly.
Anthropic Claude Models
Claude models are often praised for their natural-sounding writing and nuanced understanding of instructions.
- Default in AI Product Tools:
claude-sonnet-4-6, a balanced Claude model for nuanced product copy. - Claude Opus models: Good for premium descriptions where quality matters most.
- Claude Haiku models: Useful for simpler tasks where speed and cost matter more.
Best for: Nuanced writing, following complex brand guidelines, and natural-sounding descriptions.
OpenRouter
OpenRouter is a unified interface that gives you access to models from many different providers, including Llama, Mistral, and more.
- Default in AI Product Tools:
openrouter/free, useful for testing OpenRouter connectivity. - Custom Model Input: OpenRouter allows you to enter custom model IDs if you want to use a specific model not listed in the dropdown.
Best for: Experimenting with different models and accessing open-source alternatives.
Choosing the Right Model for Your Task
For Bulk Operations (100+ products)
Use fast models like Gemini Flash or lower-cost OpenRouter models. They complete large jobs quickly and help control provider token costs.
For High-Quality Individual Descriptions
Use a stronger model from your selected provider, such as a higher-capability GPT or Claude model. These usually take more time and cost more, but can produce better copy.
For SEO Content
Use a balanced model such as the plugin defaults or a stronger model for your top products. Always review SEO titles and descriptions before applying them.
For the Chatbot
Use faster balanced models for chatbot responses. Slower premium models can improve answer quality but may increase customer wait time and provider token costs.
Start with the plugin default model for your provider. Change models only after testing output quality, speed, and provider cost on a small product sample.
Advanced Parameters
Temperature
Temperature controls the "creativity" of the AI.
- Low Temperature (0.1 - 0.3): More consistent and factual. Good for technical specs.
- High Temperature (0.7 - 0.9): More creative and varied. Good for marketing copy.
What's Next?
If you want the AI to follow a specific pattern every time, learn about Example Descriptions.