update-ai-models
Search the web for the latest AI models from Anthropic, OpenAI, and Google, then update src/models.ts with new model IDs. Use when user asks to update or add AI models.
When & Why to Use This Skill
This Claude skill automates the maintenance of AI-powered applications by searching for the latest model releases from Anthropic, OpenAI, and Google, and programmatically updating the project's model configuration files. It ensures your codebase stays synchronized with the newest LLM capabilities and API IDs without manual documentation research.
Use Cases
- Automated Model Registry Updates: Keep your application's model list current by automatically fetching and integrating new model IDs (like the latest Claude, GPT, or Gemini versions) directly into your source code.
- Version Migration & Optimization: Quickly replace deprecated model IDs with newer, more cost-effective, or higher-performing alternatives as soon as they are released by providers.
- Type-Safe Configuration Management: Maintain TypeScript integrity by automatically updating model definitions and running verification scripts (npm run build/test) to ensure new configurations don't break the application.
| name | update-ai-models |
|---|---|
| description | Search the web for the latest AI models from Anthropic, OpenAI, and Google, then update src/models.ts with new model IDs. Use when user asks to update or add AI models. |
Update AI Models
Search for the latest AI models and update the model list in this project.
Instructions
Search for latest models from each provider:
- Anthropic Claude models (claude-opus, claude-sonnet, claude-haiku)
- OpenAI models (GPT series, o-series reasoning models)
- Google Gemini models
Get exact model IDs from official documentation:
Update src/models.ts:
- Add new models with correct
modelID strings - Keep
claude-4.5-sonnetas the first entry (default model) - Order models by recency within each provider section
- Remove deprecated models
- Add new models with correct
Run verification:
npm run typecheck npm run build npm test
Model Entry Format
'model-key': {
provider: 'anthropic' | 'openai' | 'google',
name: 'Display Name',
model: 'exact-api-model-id',
},
Examples
Anthropic Claude:
'claude-4.5-sonnet': {
provider: 'anthropic',
name: 'Claude 4.5 Sonnet',
model: 'claude-sonnet-4-5-20250929',
},
OpenAI:
'gpt-5': {
provider: 'openai',
name: 'GPT-5',
model: 'gpt-5',
},
'o3': {
provider: 'openai',
name: 'o3',
model: 'o3',
},
Google Gemini:
'gemini-2.5-pro': {
provider: 'google',
name: 'Gemini 2.5 Pro',
model: 'gemini-2.5-pro',
},
Notes
- Default model should be cost-effective (Sonnet tier, not Opus)
- Update tests in
src/models.test.tsif DEFAULT_AI_MODEL changes - Include sources in your response after updating