Model Configuration LLM

Configure AI model parameters and behavior.

Model Selection

channels:
  - name: main
    model: gpt-4o           # Default model
    models:                 # Available models
      - gpt-4o
      - gpt-4o-mini
      - gpt-4-turbo

Model Parameters

Parameter	Type	Range	Description
temperature	number	0-2	Creativity (higher = more random)
maxTokens	number	1-128000	Max output tokens
topP	number	0-1	Nucleus sampling
frequencyPenalty	number	-2 to 2	Reduce repetition
presencePenalty	number	-2 to 2	Encourage new topics

channels:
  - name: creative
    model: gpt-4o
    temperature: 1.2        # More creative
    maxTokens: 2048
    
  - name: precise
    model: gpt-4o
    temperature: 0.3        # More precise
    maxTokens: 4096

Popular Models

OpenAI

Model	Context	Best For
gpt-4o	128K	General, vision
gpt-4o-mini	128K	Fast, cheap
o1 / o3-mini	128K	Reasoning
gpt-4-turbo	128K	Complex tasks

Claude

Model	Context	Best For
claude-sonnet-4	200K	Latest, best overall
claude-3-5-sonnet	200K	Strong performance
claude-3-5-haiku	200K	Fast, cheap
claude-3-opus	200K	Complex tasks

Gemini

Model	Context	Best For
gemini-2.5-flash	1M	Latest Flash
gemini-2.5-pro	1M	Latest Pro
gemini-2.0-flash	1M	Fast, general
gemini-1.5-pro	2M	Long context

DeepSeek

Model	Context	Best For
deepseek-chat	64K	V3, Chinese & coding
deepseek-reasoner	64K	R1 reasoning

Others

Model	Provider	Best For
grok-3 / grok-3-mini	xAI	General, tools
mistral-large-latest	Mistral	Flagship
llama-3.3-70b-versatile	Groq	Ultra-fast inference
glm-4-plus / glm-4-flash	Zhipu AI	Chinese
qwen-max / qwen-plus	Qwen	Chinese

Per-Group Models

groups:
  123456789:
    channel: creative
    model: gpt-4o
  987654321:
    channel: precise
    model: gpt-4o-mini

Preset Override

# data/presets/coder.yaml
name: coder
model: gpt-4o
temperature: 0.2
maxTokens: 4096

Token Management

Token Usage

Monitor token usage in Web Panel → Statistics

context:
  maxMessages: 20        # Limit context length
  maxTokens: 8000        # Max context tokens

Next Steps

• Channels - API channels
• Context - Context management