Gemma 2 2B

by Google · gemma-2 family

2B

parameters

text-generation code-generation multilingual summarization

Gemma 2 2B is Google's compact open model built on the same research and technology behind the Gemini models. Despite its small size, it delivers impressive performance on text generation and coding tasks, benefiting from Google's extensive training data and techniques. This model is ideal for resource-constrained environments, running smoothly on laptops and even mobile devices. It supports multiple languages and provides fast inference, making it a strong choice for lightweight local AI applications.

Quick Start with Ollama

ollama run 2b-instruct-q8_0
Creator Google
Parameters 2B
Architecture transformer-decoder
Context Length 8K tokens
License Gemma Terms of Use
Released Jun 27, 2024
Ollama gemma2:2b

Quantization Options

Format File Size VRAM Required Quality Ollama Tag
Q4_K_M 1.4 GB 2.7 GB
2b-instruct-q4_K_M
Q8_0 recommended 1.8 GB 4 GB
2b-instruct-q8_0
F16 3.8 GB 6 GB
2b-instruct-fp16

Compatible Hardware for Q8_0

Showing compatibility for the recommended quantization (Q8_0, 4 GB VRAM).

Benchmark Scores

51.3
mmlu