Supported LLM Base Models

Navigator supports a variety of open source LLMs, and is adding more all the time. Below is the current list of supported models with links to their Hugging Face model pages where you can learn more about each model's strengths and weaknesses.

For most use cases that require training or inference on consumer hardware, we recommend using a 7B parameter model. This is the sweet spot between model performance and size for consumer devices. All LLM features in Navigator work well with 7B parameter models.

Full List of Supported Models

• CodeGemma Models
  • CodeGemma 7B IT 8bit
• Codestral Models
  • Codestral 22B v0.1 4bit
  • Codestral 22B v0.1 8bit
• DeepSeek Models
  • DeepSeek V3 0324
  • DeepSeek V3 0324 8bit
  • DeepSeek V3 0324 4bit
  • DeepSeek R1 Distill Llama 8B
  • DeepSeek R1 4bit
• Gemma Models
  • Quantized Gemma 7B IT
  • Gemma 3 1B IT 4bit
  • Gemma 3 1B IT 8bit
  • Gemma 3 1B IT 6bit
  • Gemma 3 1B IT bf16
  • Gemma 3 Text 12B IT 4bit
  • Gemma 3 Text 27B IT 4bit
  • Gemma 3 4B IT 4bit
  • Gemma 3 4B IT 6bit
  • Gemma 3 4B IT bf16
• Llama Models
  • Meta Llama 3.1 8B Instruct 4bit
  • Meta Llama 3.1 8B Instruct 8bit
  • Meta Llama 3.1 70B Instruct 4bit
  • Llama 3.1 Nemotron 70B Instruct HF 4bit
  • Llama 3.3 70B Instruct
  • Meta Llama 4 Scout 17B 16E bf16
  • Meta Llama 4 Scout 17B 16E fp16
  • Llama 4 Scout 17B 16E Instruct
  • Llama 4 Scout 17B 16E
  • Llama 4 Maverick 17B 128E Instruct
  • Llama 4 Maverick 17B 128E Instruct FP8
  • Llama 4 Maverick 17B 128E
  • Llama 4 Maverick 17B 16E Instruct 6bit
  • Llama 4 Maverick 17B 16E Instruct 4bit
  • Llama 3 Taiwan 70B Instruct
  • Llama 3.1 Nemotron Nano 8B v1
• Ministral Models
  • Ministral 8B Instruct 2410 bf16
• Mistral Models
  • Mistral 7B Instruct v0.3
  • Mistral NeMo Minitron 8B Instruct
• Mixtral Models
  • Mixtral 8x7B Instruct v0.1
  • Mixtral 8x22B Instruct v0.1
• Phi Models
  • Phi 3 Medium 128K Instruct bf16
  • Phi 3.5 Mini Instruct bf16
  • Phi 4
• Qwen Models
  • Qwen2 7B Instruct 8bit
  • Qwen2 72B Instruct 4bit
  • Qwen2.5 72B Instruct 4bit
  • Qwen2.5 14B Instruct
  • Qwen2.5 Coder 7B Instruct
  • QwQ 32B Preview
• Qwen3 Models
  • Qwen3 0.6B 3bit
  • Qwen3 0.6B 4bit
  • Qwen3 0.6B 6bit
  • Qwen3 0.6B 8bit
  • Qwen3 0.6B bf16
  • Qwen3 1.7B 3bit
  • Qwen3 1.7B 4bit
  • Qwen3 1.7B 6bit
  • Qwen3 1.7B 8bit
  • Qwen3 1.7B bf16
  • Qwen3 4B 3bit
  • Qwen3 4B 4bit
  • Qwen3 4B 6bit
  • Qwen3 4B 8bit
  • Qwen3 4B bf16
  • Qwen3 8B 3bit
  • Qwen3 8B 4bit
  • Qwen3 8B 6bit
  • Qwen3 0.6B
  • Qwen3 1.7B
  • Qwen3 4B
  • Qwen3 8B
  • Qwen3 30B A3B 4bit
  • Qwen3 30B A3B 6bit
  • Qwen3 30B A3B 8bit
  • Qwen3 30B A3B
  • Qwen3 235B A22B

Model Selection Considerations

When selecting a model, consider these key factors:

• Parameter Count: Generally, larger models (higher parameter counts) offer better performance but require more computing resources.
• Memory Requirements: Each model requires a specific amount of RAM. Ensure your hardware meets these requirements.
• Specialization: Some models are optimized for specific tasks (e.g., code generation, instruction following).
• Quantization: Many models offer quantized versions (4-bit, 8-bit) that reduce memory requirements at a small cost to performance.