AI Model Pricing

AI Model Pricing defines the per-model input and output prices that the platform uses to calculate AI usage costs for all tenants. System administrators maintain these entries as AI providers update their pricing or introduce new models. Each entry maps a provider, a model-name pattern, and a match mode to USD prices per million tokens.

Accessing the Page

• Route: /ai-model-pricing
• Menu Path: Settings → System → AI Model Pricing
• Primary audience: System administrators only. These pages are accessed via the system admin login at system.portal.net — not the tenant portal.

What you can do here

• View all configured AI model pricing entries with provider, model pattern, match mode, display name, input price, output price, and active status.
• Add a new pricing entry for a provider and model pattern.
• Edit an existing entry to update prices, the display name, match mode, or active state.
• Reorder entries with drag handles to change matching precedence.
• Delete a pricing entry that is no longer needed.
• Deactivate an entry by toggling its Active flag off — inactive entries are excluded from cost calculations without being deleted.

Common tasks

1. Open Settings → System → AI Model Pricing at system.portal.net.
2. Click Add Pricing to open the add dialog.
3. Select the provider, enter the model pattern, choose the match mode, and set the display name plus input and output prices per million tokens.
4. Save the entry and verify it appears in the list with the correct status.
5. Drag entries into the desired order if you need to change which pattern takes precedence.
6. After updating prices, tenant AI budget calculations will use the new values for subsequent usage records.

Notes

• Model Pattern is matched against the model name string returned by the AI provider. Use Substring for partial matching (e.g., claude-opus) or Regex for exact and pattern-based matching.
• Prices are expressed in USD per one million tokens. Use the value shown in the AI provider’s pricing documentation.
• The on-screen order controls priority when multiple entries could match the same model name. Move entries with the drag handle to change precedence.
• Providers now include Anthropic, OpenAI, Google, Mistral, Cohere, Meta, xAI, DeepSeek, OpenRouter, and Qwen.
• Inactive entries are excluded from cost calculations but remain in the list for reference. Use the Active toggle to disable an entry without losing its configuration.
• Deleting an entry is permanent and immediate. Historical usage records already calculated against the deleted entry are not affected.
• The charged prices (shown to tenants) may include a markup applied at the platform level on top of the base prices you configure here.

Adding a Pricing Entry

Click Add Pricing in the page header. A dialog opens where you configure the new entry.

Fields

Field	Required	Type	Description	Default	Validation
Provider	Yes	Select	The AI provider: Anthropic, OpenAI, Google, Mistral, Cohere, Meta, xAI, DeepSeek, OpenRouter, or Qwen	Anthropic	Required
Model Pattern	Yes	Text	String matched against the provider’s model name (e.g., claude-opus, gpt-4o)	—	Required; matched against provider model name strings
Match Mode	Yes	Select	Matching strategy used for the model pattern	Substring	Required; choose Substring or Regex
Display Name	Yes	Text	Friendly label shown in budget and usage pages (e.g., Claude Opus)	—	Required
Input Price ($/1M tokens)	Yes	Number	Cost per one million input tokens, in USD	0.00	Must be ≥ 0
Output Price ($/1M tokens)	Yes	Number	Cost per one million output tokens, in USD	0.00	Must be ≥ 0
Active	—	Toggle	Whether this entry is used in cost calculations	On	—

Steps:

1. Click Add Pricing in the page header.
2. Select the Provider from the dropdown.
3. Enter the Model Pattern — use the model name or a pattern documented by the provider.
4. Choose the Match Mode to control whether the pattern is treated as a substring or a regular expression.
5. Enter a human-readable Display Name that will appear in tenant budget and usage pages.
6. Enter the Input Price and Output Price in USD per million tokens. Use the provider’s current published pricing.
7. Leave Active on unless you want the entry inactive at creation.
8. Click Add Pricing to save.

Editing a Pricing Entry

Click the Edit button (pencil icon) on any row to open the edit dialog with the entry’s current values pre-filled.

Same fields as Adding a Pricing Entry, except the provider is locked after creation. Use drag-and-drop on the main list page to change the order of entries.

Steps:

1. Click the Edit button on the target row.
2. Update the desired fields (e.g., adjust prices when the provider changes its rates or switch the match mode).
3. Click Save Changes.

Deleting a Pricing Entry

Click the Delete button (trash icon) on any row.

1. Click the Delete button on the target row.
2. Read the confirmation: “Are you sure you want to delete the pricing for [model]? This action cannot be undone.”
3. Click Delete to confirm — the entry is permanently removed from the list.

Note: If you want to stop using an entry without permanently deleting it, use the Active toggle in the edit dialog to deactivate it instead.

Related Pages

• AI Settings — tenant-level AI configuration including budgets, credentials, and usage monitoring
• Tenant Management — tenant organizations whose AI usage costs are calculated using these pricing entries