Gemma
by Google DeepMind · ai.google.dev ↗
Google's family of lightweight open-weight models derived from the same research as Gemini.
llm open-weights lightweight
- Category
- Open Models
- Business model
- Open (open weights/source)
- Availability
- Global
- Launched
- 2024-02
- Record updated
- 2026-07-27
- Model family
- Gemma
- License
- Gemma Terms of Use
- Founded
- 2010
- Headquarters
- GB
- Canonical URL
- https://globalaiproductindex.com/products/gemma/
Overview
Gemma is Google's family of lightweight open-weight models, derived from the same research that produced Gemini. The models target laptops, workstations, and single-GPU servers, with variants for vision, code, and on-device use. Weights are freely available under Google's Gemma license.
Key features
- Compact open-weight models from sub-billion to mid-size scales
- Multimodal and code-focused variants
- Optimized for consumer hardware and on-device inference
- Supported across Ollama, Hugging Face, and Google AI tooling
Use cases
- Running capable models locally on consumer hardware
- On-device and edge AI applications
- Fine-tuning small models for focused tasks
Pricing
Gemma weights are free to download under the Gemma license; hosted inference is available through Google and third parties at usage prices.
Frequently asked questions
Is Gemma open source?
Gemma models are released with open weights under Google's Gemma license, permitting broad use including commercial applications.
How is Gemma related to Gemini?
Gemma is built from the same research and technology as Google's Gemini models, in smaller open-weight form.
Can Gemma run locally?
Yes. Gemma is designed for laptops and single-GPU machines and runs in tools like Ollama and LM Studio.
Similar products
- Baichuan — Open-weight bilingual LLM family from Beijing-based Baichuan Intelligence, now focused on medical AI with its Baichuan-M reasoning models.
- Command R — Cohere's open-weight LLM family optimized for retrieval-augmented generation and tool use; succeeded by Command A as Cohere's flagship.
- DBRX — Open-weight mixture-of-experts LLM released by Databricks in 2024; retired from Databricks serving in 2025, though the weights remain available.
- DeepSeek — Open-weight LLM family and chat assistant from Hangzhou-based DeepSeek, known for strong reasoning models. The official DeepSeek-V4-Flash-0731 API brings a major agent-capability upgrade via post-training, with native Responses API support and Codex adaptation.
- Falcon — Family of open-weight large language models developed by the Technology Innovation Institute (TII).
- GLM — Family of open-weight large language models from Zhipu AI, spanning general, reasoning, and coding-focused variants.
Sources
This record was last reviewed on 2026-07-27.
Machine-readable record: /api/products/gemma.json · Spot an error? Suggest a correction