Models

Document Information

Version: v1.0 Updated: 2026-08-31

Feature Overview

Item	Content
Applicable Roles	Model Provider, Model Consumer
Navigation Path	Model Services > Discover > Models
Page Route	/modelone/store/model
Managed Objects	Model listings, model details, provider instances, and quick-start information

Beginner Explanation

Models works like a searchable model catalog. Find a model with the required input and output capabilities, review its provider, price, context, and status, and then decide whether to try it or integrate it from the quick-start information.

Terminology

Term	Description
Model Author	The organization or brand associated with a model.
Model Source	The upstream service that provides a callable model instance.
Context Length	The total input and output capacity of one request.
Quick Start	The integration area that shows the URL, protocol, and authentication method.
Personal Key	A personal credential that a Model Consumer uses to call a model through the platform.
API Key	A general term for an API credential. Use a Personal Key for the examples on this page. A Model Provider can use a separate upstream API Key when publishing a model.
Request header	The fields sent with a request. A request header commonly carries authentication data.
Endpoint	The target URL that receives a request. It combines the Base URL and API path.
Model ID	The identifier that selects the target model instance for a call.

First-Use Paths

Role	First-Use Path	Completion Signal
Model Consumer	Prepare a Personal Key → select a model and provider in "Models" → make one controlled call in "Playground" → find the request in "My Calls" → review consumption in "Model Usage"	The response area shows a response or error, Call Logs contains the request, and Model Usage contains the related record.
Model Provider	Create a model in "My Models" → publish it from "My Deployments" → review calls in "Customer Calls" → review revenue in "Model Earnings"	The model and deployment appear in their lists, and the customer-call and earnings pages show the related statistics.

When a model has multiple providers, compare price, status, and context first. A model in the list is not proof of a successful call. Confirm the response, Call Logs record, and Model Usage record for the same request.

Prerequisites

1. The current account can open "Models".
2. The target model has been listed and is visible to the current account or customer.
3. Before calling, quota, pricing, context limits, and terms of use have been confirmed.

Call and Billing Risk

Trying a model, submitting a prompt, or calling an API creates call records and may consume credits or generate billing records. Before calling, verify the model, provider, price, quota, and usage scope.

Page Description

The page provides model search, capability filters, sorting, and listings. Open a model to use the Providers, Quick Start, Performance, and Overview tabs. Provider cards also provide Try Now and Quick Start entries.

Page screenshots:

Focus on the search field, capability filters, and model list. The active filters jointly determine the result set.

Main Operations

Query Models

1. Go to Model Services > Discover > Models.
2. Enter a model name, author, series, or source in the search field.
3. Narrow the results by input capability, output capability, context, billing, model author, model source, or scenario.
4. Verify the model name, capability labels, and status. Clear combined filters and query again if no result is returned.

View Model Details

1. In the model list, click the target model card to enter its details page.
2. Verify the model name, author, type, input and output capabilities, context length, and billing method.
3. Confirm the model summary, capability labels, modalities, and protocol information displayed in the header.
4. The page body provides four tabs: Providers, Quick Start, Performance, and Overview. Switch between them based on your follow-up task.

Compare Provider Instances

1. Select the "Providers" tab on the model details page.
2. Compare provider instance status, price, context, latency, throughput, and success rate.
3. Before trying or integrating the model, confirm the Model ID, protocol, and status on the target provider card.

Verify the pricing, maximum context, latency, and success rate metrics for each provider instance to determine the target instance.

Try a Model

1. Click "Try Now" on the target provider card.
2. The platform opens the Playground page corresponding to the model type, such as text chat, image generation, video generation, or speech generation.
3. In Playground, verify that the model and provider shown at the top match the selection on the details page.
4. Before trying the model, confirm the Personal Key, price, and quota. Submitting a prompt or generating content creates call records and consumes usage.

View Quick Start

1. Select the "Quick Start" tab, or click "Quick Start" on the target provider card.
2. Confirm the selected provider and switch the code example between SDK, HTTP, and Curl.
3. Review the Model ID, Base URL, Path, Full URL, request headers, and authentication method.
4. Record integration information with placeholders such as <BASE_URL>, <ENDPOINT_PATH>, <API_KEY>, or <PERSONAL_KEY>. Do not store real credentials.

Inspect the integration code sample in the Quick Start tab to verify the Base URL, request path, and authentication header format.

View Performance

1. Select the "Performance" tab on the model details page.
2. Set the time range and data granularity to review.
3. Review average request latency, average first-token latency, real-time request frequency, and request success rate.
4. When comparing time ranges or providers, keep the filters consistent so that metrics use the same scope.

Review average response latency, first-token latency, and request success rate trends under the selected time range and granularity.

View Model Overview

1. Select the "Overview" tab on the model details page.
2. Review the model description, capability tags, input/output modalities, context, and usage boundaries.
3. Keep the overview consistent with the provider instance and the Model ID and protocol shown in Quick Start.

Review the capability description, supported modalities, and parameter boundaries in the Overview tab to ensure model characteristics meet business requirements.

Parameter Reference

Field Name	Required	Field Type	Example	Description
Model Name	Yes	Text	Example Model	Display name used to identify the model in the list and details page.
Provider	Yes	Text / Filter	Example Provider	Tenant or channel that provides the model instance.
Model Type	No	Filter / Tag	Text	Distinguishes multimodal, text, image, speech, video, embedding, reranking, and other model types.
View	No	Switch	Table / Card	Changes the presentation without changing model visibility.
Capability Tags	No	Tag	Tool Calling	Shows model capabilities such as tool calling or reasoning.
Input/Output Modalities	No	Tag	Text / Image	Shows supported input and output types.
Billing Method	No	Text	Credits / 1M Tokens	Shows the model billing unit for input, output, or per-request pricing.
Status	No	Tag	Published	Shows the status of the model or provider instance.
Actions	No	Button	View, Try Now, Quick Start	Used to open model details, enter Playground, or view integration information.

Pitfalls

• The same model may have multiple providers. Confirm the selected provider instance before calling.
• Copy the exact Model ID from the provider instance.
• Replace the API Key placeholder in each quick-start example with an authorized Personal Key.
• Trying a model or submitting a prompt creates call records and can consume credits or generate billing records. Before submission, verify the model, Personal Key, input, and expected usage.

Result Validation

Check Item	Success Signal	If Abnormal
Model list is accessible	Model cards or list entries are displayed, and the filter area loads correctly.	Check account permissions, navigation path, and page loading status.
Filters can be selected	After changing model type, capability, provider, or scenario filters, list results update accordingly.	Clear filters and search again, or refresh the page.
Model details can be opened	Clicking View opens the details page with model introduction, providers, pricing, context, and modality information.	Return to the list and open the model again, or check whether the model is still visible to the current account.
Details match the list	Model name, author, status, and billing information in details are consistent with the list.	Use the details page as the source of truth and report synchronization issues if needed.
Try Now entry is visible	Clicking Try Now opens the Playground page for the corresponding modality, and the model and provider at the top match the details page.	Return to model details, select a listed provider, and verify model type and Playground permission.
Quick Start opens	The Quick Start tab or provider-card entry shows the protocol, Model ID, URL, and authentication method.	Verify provider instance status, then check call permission and visibility for the current account.
Performance data is visible	The Performance tab shows metrics for the selected time range and data granularity.	Adjust the time range and granularity, and verify that the selected provider has call data.
Overview tab opens	The Overview tab shows the model description, capabilities, modalities, and usage boundaries.	Refresh the model details page and ask the Model Provider to complete missing fields.

FAQ

Cannot Find the Target Model

Symptom:

The target model does not appear after a name or capability query.

Possible Causes:

• Combined filters are too narrow.
• The model is not visible to the current account.

Resolution:

1. Clear filters and query again.
2. Verify model visibility and provider status.

Model Details Are Incomplete

Symptom:

Capabilities, context, or billing information is missing.

Possible Causes:

• Model information is incomplete.
• Page data has not synchronized yet.

Resolution:

1. Refresh details and review the provider tab.
2. Do not integrate a model with missing usage boundaries.

No Provider Is Listed

Symptom:

No provider instance can be selected in model details.

Possible Causes:

• Provider instances are not listed or are disabled.
• The current account is outside the authorization scope.

Resolution:

1. Verify instance status and authorization.
2. Ask the Model Provider when an instance will be listed.

Billing or Context Is Unclear

Symptom:

Price or context differs across displayed records.

Possible Causes:

• The records use different provider instances.
• Billing units or model versions differ.

Resolution:

1. Compare the same provider and version.
2. Record the scope shown in current model details.

Quick Start Is Unavailable

Symptom:

Quick Start does not open or lacks integration information.

Possible Causes:

• The model instance is not ready for calls.
• The account lacks call permission.

Resolution:

1. Verify model and provider status.
2. Check call authorization and Personal Key.

Notes

• Do not externally share screenshots containing call credentials or customer identifiers from details pages.
• Model capabilities, pricing, and context limits are subject to the details page.
• Replace the Endpoint and API Key placeholders in quick-start examples with values from your tenant.

Next Steps

1. Compare provider instances in model details and review Model ID, context length, pricing, rate limits, and input/output modalities.
2. Click "Try Now" to enter the corresponding Playground page, or click "Quick Start" to obtain a redacted integration example.
3. Review Performance and Overview to confirm latency, success rate, capabilities, and usage boundaries.
4. Before production integration, confirm quota, call limits, and Model Source status.