Model Selection¶

QUASAR supports multiple AI models from different providers.

Auto Mode (Default)¶

By default, QUASAR automatically selects the best model:

quasar "your request"

Priority Order:

Cerebras - Best for tool calling
Ollama - Local fallback
Groq - Cloud fallback

Manual Selection¶

Use --model to specify a model:

quasar --model provider/model-name "your request"

Examples¶

# Cerebras
quasar --model cerebras/zai-glm-4.7 "create a hello.py"

# Groq
quasar --model groq/openai/gpt-oss-120b "explain main.py"

# Ollama (local)
quasar --model ollama/qwen2.5-coder:7b "fix the bug"

Available Models¶

Cerebras¶

Model	Description
`zai-glm-4.7`	Default, balanced
`qwen-3-235b-a22b-instruct-2507`	Large, powerful

Groq¶

Model	Description
`openai/gpt-oss-120b`	Best for tools
`openai/gpt-oss-20b`	Smaller, faster
`llama-3.3-70b-versatile`	General purpose

Ollama¶

Model	Description
`glm-4.7:cloud`	Default, balanced
`deepseek-v3.1:671b-cloud`	Code intelligence
`gpt-oss:120b-cloud`	Large, powerful
`qwen3-coder:480b-cloud`	Coding focused

Custom Models

Use --model ollama/your-model for any local model.

Fallback Behavior¶

When using Auto mode:

Tries primary model
If fails (rate limit, error), tries next
Continues until success or all fail

When using --model:

Uses only specified model
No fallback to others
Fails if model unavailable

Tips¶

Best for Tool Calling

Use cerebras/zai-glm-4.7 or groq/openai/gpt-oss-120b

For Privacy

Use ollama/ models for local processing

For Speed

Groq models are fastest for inference