1. Provider: Filter by official source — for example, select Google, OpenAI, or Anthropic to instantly view all available models from that vendor.
2. Tags: Filter by model capability — for example, select Multimodal or Image for image processing, or Discounted or Value for budget-friendly options.
3. Available Token Groups: This determines which server or channel group is used when calling the model. Different groups may correspond to different levels of stability and pricing tiers. Choose based on your specific requirements. For detailed information, see: API Key 4. Model Name: The English text in the upper-left corner of the card is the standard model identifier. When writing code or configuring third-party clients, this name must be entered exactly.
5. One-Click Copy: Click the [copy icon] next to the name to copy it instantly. Strongly recommended to avoid manual input errors.
6. Feature Tags: Tags in the lower-right corner of the card (e.g., Text, Popular) indicate the model's primary use case and popularity.
7. Multiplier Information (Core Billing Rules):Model: This is the "input multiplier" (Prompt). A value of Model: 1 means the model charges at 1x the base rate ($1 per 500K tokens) for processing your prompt.
Completion: This is the "output multiplier" (Completion). A value of Completion: 6 means the model charges at 6x the input rate for generating responses.
8. Promotional Discount: Shows the current promotional pricing. For example, Discount: 30% means your actual cost for calling this model is only 30% of the original price!