Model List

The AIClient-2-API supports listing models provided by backend LLMs. Multiple providers are supported, and the default model list is defined in src/providers/provider-models.js.

OpenAI Compatible Interface

GET http://localhost:3000/v1/models

This returns a response mimicking the OpenAI models endpoint, containing the list of LLM models configured in the backend.

Gemini Compatible Interface

GET http://localhost:3000/v1beta/models

This returns a response mimicking the Gemini models endpoint.

---

Supported Models Preview

Below is a preview of pre-defined models for major providers (note: these can be dynamically extended via configuration):

Google Gemini (GCP OAuth)

• gemini-2.5-flash
• gemini-2.5-pro
• gemini-3.1-pro-preview
• gemini-3.1-flash-lite-preview
• gemini-3-pro-preview
• gemini-3-flash-preview

Anthropic Claude (Kiro/CodeWhisperer)

• claude-3-7-sonnet-20250219
• claude-sonnet-4-6
• claude-opus-4-6
• claude-haiku-4-5

Grok (xAI)

• grok-3
• grok-3-mini
• grok-3-thinking
• grok-4
• grok-4-mini
• grok-4-thinking

OpenAI Codex

• gpt-5
• gpt-5-codex
• gpt-5.4

Qwen / iFlow

• qwen3-coder-plus
• qwen3-max
• deepseek-v3.2
• deepseek-r1
• kimi-k2.5
• glm-5

---

Automatic Model Selection (AUTO Mode)

If MODEL_PROVIDER is set to auto in the configuration, the system automatically matches the best provider based on the model name in the request:

1. The system iterates through all configured provider pools.
2. Matches the model name with the appropriate driver.
3. If the primary account fails, it automatically switches to a backup account within the same or a compatible protocol.