Add Lifecycle Status to AI Models in the Developer Portal
Background
The Developer Portal currently lists all available AI models but does not communicate their lifecycle stage.
To better communicate expectations around stability, availability, and long-term planning, each model should expose its current lifecycle status.
The Developer Portal should become the single source of truth for model lifecycle information.
Implementation
- Add a new Lifecycle column to the AI model overview.
- Display the lifecycle status for each model.
- Support the following lifecycle stages:
- Experimental
- Testing
- Stable
- Long-Term Support (LTS)
- Add documentation describing each lifecycle stage.
Current Lifecycle
| Model |
Lifecycle |
| gpt-oss-120b |
LTS |
| Qwen3.5-0.8B |
Stable |
| Ministral-3-14B-Instruct-2512 |
LTS |
| Mistral-Medium-3.5-128B |
Testing |
| Qwen3.5-122B-A10B-FP8 |
Stable |
| Qwen3.6-35B-A3B-FP8 |
Stable |
| GLM-OCR |
Testing |
| Qwen3-Embedding-8B |
LTS |
| Qwen3-VL-Reranker-2B |
Testing |
| whisper-large-v3-turbo |
LTS |
Lifecycle Definitions
🧪 Experimental
Early preview intended for exploration and feedback.
- Features and behavior may change at any time.
- The model may be replaced or removed without notice.
- No guarantees regarding stability or availability.
- Not recommended for production workloads.
🔬 Testing
Available for development, evaluation, and early integrations.
- The model is being validated under real-world conditions.
- Occasional downtime, performance fluctuations, or service interruptions may occur.
- Not recommended for production workloads.
✅ Stable
Recommended for production use.
- Covered by normal operational processes.
- At least 90 days' notice before deprecation or removal.
- Recommended replacement models will be provided.
- May be promoted to Long-Term Support (LTS) after proving itself over time.
🛡️ Long-Term Support (LTS)
Recommended for long-term production deployments.
- Previously proven as a Stable model.
- Extended lifecycle commitment.
- At least 12 months' notice before deprecation or removal.
- Recommended replacement models will be provided.
Lifecycle Flow
Experimental
↓
Testing
↓
Stable
↓
Long-Term Support (LTS)
Not every model will progress through every stage. Models may be removed during the Experimental or Testing phase, or remain Stable without ever becoming LTS.
Add Lifecycle Status to AI Models in the Developer Portal
Background
The Developer Portal currently lists all available AI models but does not communicate their lifecycle stage.
To better communicate expectations around stability, availability, and long-term planning, each model should expose its current lifecycle status.
The Developer Portal should become the single source of truth for model lifecycle information.
Implementation
Current Lifecycle
Lifecycle Definitions
🧪 Experimental
Early preview intended for exploration and feedback.
🔬 Testing
Available for development, evaluation, and early integrations.
✅ Stable
Recommended for production use.
🛡️ Long-Term Support (LTS)
Recommended for long-term production deployments.
Lifecycle Flow
Not every model will progress through every stage. Models may be removed during the Experimental or Testing phase, or remain Stable without ever becoming LTS.