Scope of Model Support
To ensure the stability and maintainability of the SERVICEME platform in a multi-cloud and multi-model environment, we have established a unified support standard for the integrated large language models (LLM), embedding models, speech and document recognition models.
This standard is intended to:
- Reduce the complexity of multi-model compatibility and operations
- Improve overall performance and controllability
- Ensure the stable availability of new features on mainstream models
Model Support Levels
Standard Support
- Definition: Fully included in the product support scope.
- Requirements: Functionality, performance, and stability have all completed full testing, and service availability assurance (SLA, Service Level Agreement) is provided.
Experimental Support
- Definition: The model has been integrated into the main workflow to ensure that the functionality is available.
- Requirements: Only basic functional testing is performed (such as interface connectivity and successful execution of the main workflow), while performance/stability validation is not yet completed.
- Identification: Clearly marked as "experimental" in the product documentation.
Model Categories and Support Status
Large Language Model (LLM)
| Interface/Vendor | Supported Models | Current Support Level | Notes |
|---|
| Azure OpenAI | GPT-5.4, GPT-5.5 | Standard Support | Deep Research/Super Agent/Ai Slides requires disabling reasoning mode |
| OpenAI | GPT-5.4, Qwen3.6-Max-Preview, Qwen3.6-Plus, Qwen3.6-Flash, DeepSeek V4 | Standard Support | Deep Research/Super Agent/Ai Slides requires disabling reasoning mode |
| Google Vertex | Gemini 3.1 Pro, Gemini 3.1 Flash | Standard Support | Does not support AI Translator |
Text-to-Image Models
| Vendor | Supported Models | Current Support Level | Notes |
|---|
| Azure | GPT-Image-1.5, GPT-Image-2 | Standard Support | |
Embedding Models
| Vendor | Supported Models | Current Support Level | Notes |
|---|
| Azure | text-embedding-3-large | Standard Support | |
Image Recognition / OCR
| Vendor | Supported Models | Current Support Level | Notes |
|---|
| Azure | Azure AI Document Intelligence | Standard Support | |
| Alibaba Cloud | Alibaba Cloud OCR | Standard Support | |
| Third-party | TextIn | Standard Support | |
Speech Recognition (STT)
| Vendor | Supported Models | Current Support Level | Notes |
|---|
| Azure | Whisper | Standard Support | |
Data Sources (Data Source)
| Type | Example | Current Support Level | Notes |
|---|
| MySQL | | Standard Support | |
| SQLServer | | Standard Support | |
| Azure Databricks | | Standard Support | |
| Snowflake | | Standard Support | |
| Azure Fabric | | Standard Support | |
| PostgreSQL | | Standard Support | |
Search Engines (Search Engine)
| Type | Supported Models | Current Support Level | Notes |
|---|
| Standard Support | G-bing Search, Meta Search, Tencent Search, Google Search, Aliyun Qurak Search | Standard Support | |
Access and Validation Requirements
-
Unified Interface Standards
- All models must be called through the unified Model Gateway service.
- Request and response formats must comply with the SERVICEME model interface specification.
-
Security and Compliance
- Calling third-party public APIs without security validation is prohibited.
- Integrated models must comply with company data compliance requirements (privacy, security, regional restrictions, etc.).
-
Performance and Observability
- Models must support invocation logging and performance monitoring.
- Health checks and circuit breaker mechanisms must be established for standard support models.