Local LLM model page

MiniCPM-V (8B)

Compact multimodal VLM. Strong image understanding in a small package. 3.6M downloads.

Find the best model for my hardwareBrowse all 183 LLMs
Parameters
8B
Minimum RAM
8 GB
Model size
5 GB
Quantization
Q4_K_M

Can MiniCPM-V (8B) run locally?

MiniCPM-V (8B) is best suited for entry-level laptops and desktops. LocalClaw recommends Q4_K_M as the default quantization, with at least 8 GB RAM.

Search term for LM Studio or compatible runtimes: minicpm-v-2.6

Hugging Face repository: openbmb/MiniCPM-V-2_6-GGUF

visionstandard

Strengths

  • Compact VLM
  • Strong image understanding
  • 3.6M downloads
  • Apache 2.0

Limitations

  • Limited context
  • Weak at coding
  • English/Chinese only

Best use cases

  • Image description
  • Document scanning
  • Visual Q&A
  • OCR

Benchmarks

Speed: 7/10

Quality: 7/10

Coding: 4/10

Reasoning: 6/10

Technical details

Developer: OpenBMB (Tsinghua)

License: Apache 2.0

Context window: 8,192 tokens

Architecture: SigLIP vision encoder + MiniCPM language model

Released: 2024-08

Similar models

llava-13bllava-llama3-8bmoondream2