RUN THIS LLM
Search local LLM hardware requirements
Gemma 3 27B
Google · 27B · Vision
Gemma's largest model. Strong multimodal capabilities competitive with much bigger models.
VRAM Requirements
| Quantization | VRAM |
|---|---|
| Q4_K_M (smallest) | 16.2 GB |
| Q8_0 (balanced) | 29.7 GB |
| FP16 (full quality) | 54 GB |
Specifications
- Parameters: 27B
- Category: Vision
- Max context: 128K tokens
- System RAM: 24 GB minimum
- HuggingFace: google/gemma-3-27b-it
Benchmarks
- MMLU: 79 — General knowledge & reasoning (5-shot)
- HumanEval: 72 — Code generation (pass@1)
- MATH: 56 — Competition-level math reasoning
Loading interactive analysis...