MedGemma 4B

Overview

Google's 4B medical variant of Gemma with vision and text, tuned for radiology, clinical imaging, and report drafting. 128k context, Gemma license.

When to pick this model

• Drafting or summarizing radiology and clinical reports
• Prototyping medical imaging assistants on a single small GPU
• Research into multimodal clinical NLP without API costs
• Edge deployments in healthcare workflows where data can't leave the device

VRAM requirements by quantization

Quantization	VRAM required
Q4_K_M (recommended)	2.3 GB
Q5_K_M	2.8 GB
Q8_0	4.3 GB
FP16 (no quantization)	8 GB

VRAM figures include model weights plus a typical 8k KV cache and ~600 MB runtime overhead (Ollama / llama.cpp baseline). Add headroom for higher context lengths.

Strengths

• Domain-tuned on clinical literature and radiology imagery
• Compact 4B footprint (~2.3GB VRAM at Q4)
• True multimodal — text plus medical images
• Permissive Gemma license for research and most commercial use

Limitations

• Decision-support only — not approved for direct clinical use
• Narrow specialization; weak outside medical contexts
• Gated on Hugging Face

Architecture & training

Architecture: Gemma 4 · 4B parameters · multimodal text + image · 128k context

Training: Medical variant of Gemma, fine-tuned on clinical literature, radiology imaging, and medical reports.

Quick start

ollama run medgemma

Or use the open-source MCP server to query this model from Claude Desktop, Cursor, or any MCP-compatible client.