Local LLM model page

Phi-3 Medium (14B)

Microsoft's medium Phi-3. Strong reasoning capabilities for its size. Good balance of speed and quality.

Parameters
14B
Minimum RAM
12 GB
Model size
8 GB
Quantization
Q5_K_M

Can Phi-3 Medium (14B) run locally?

Phi-3 Medium (14B) is best suited for mainstream Macs and PCs with 16 GB RAM. LocalClaw recommends Q5_K_M as the default quantization, with at least 12 GB RAM.

Search term for LM Studio or compatible runtimes: phi-3-medium-4k-instruct

Hugging Face repository: lmstudio-community/Phi-3-medium-4k-instruct-GGUF

chatreasoningpower

Strengths

  • Microsoft's medium Phi-3. Strong reasoning capabilities for its size. Good balance of speed and quality.

Limitations

  • Performance depends heavily on quantization, RAM bandwidth and runtime support.

Best use cases

  • chat
  • reasoning
  • power

Benchmarks

Speed: 6/10

Quality: 7/10

Coding: 7/10

Reasoning: 8/10

Technical details

Developer: phi

License: See model repository

Context window: Unknown tokens

Architecture: See model card

Released: 2024-04