LLM
GA
Open weights
Gemma 3 27B
Open model with image input on a single 24 GB GPU
Release2025-03-12
Context131K
Max output—
Parameters27B
Local / VRAM16–24 GB
LicenseOpen weights
Overview
Gemma 3 27B is a model in the LLM category from Google DeepMind. It was released in 2025-03-12. The vendor is based in US/UK. Best for: Open model with image input on a single 24 GB GPU.
Strengths
- Image input
- Single 4090
- Wide fine-tuning support
Limits
- Overtaken by Gemma 4
- 131K context
Specifications
| Vendor | Google DeepMind |
|---|---|
| Category | LLM |
| Release | 2025-03-12 |
| Status | GA |
| License | Open weights |
| License name | Gemma License |
| Commercial use | Allowed |
| Parameters | 27B |
| Architecture | dense transformer |
| Context window | 131K |
| Input modalities | Text, Image |
| Output modalities | Text |
| Prompt types | Text |
| Negative prompts | No |
| Max outputs | 1 |
| Local deployment | Yes |
| VRAM | 16–24 GB |
| System RAM | 32–64 GB |
| API access | Yes |
| Pricing | Free weights |
| Confidence | High |
| Vendor country | US/UK |
| Vendor type | Frontier lab |
Pricing & hardware
Free weights
This model can be self-hosted or run locally. Approximate GPU memory: 16–24 GB. Approximate system RAM: 32–64 GB. API access is available.