Gemma 3 — local AI model by Google
Written by Jakub Rusinowski · Last updated
Gemma 3 delivers exceptional quality-per-VRAM ratio, with the 4B model outperforming many 7B competitors. Features multimodal capabilities and a 128k context window across all sizes. Previous generation — superseded by Gemma 4 (E2B/E4B/26B-A4B/31B and the 12B Unified follow-up), which improves efficiency and multimodal support. Still widely used and supported.
Variants
The smallest Gemma 3 variant needs about 1 GB of VRAM at Q4_K_M — quantized weights plus framework overhead, before any KV cache.
| Model | VRAM at Q4 | VRAM | Context | Run it |
|---|---|---|---|---|
| Gemma 3 1B Instruct → 1B | ~1.4 GB | 32,000 | ollama run gemma3:1b | |
| Gemma 3 4B Instruct → 4B | ~3.2 GB | 128,000 | ollama run gemma3:4b | |
| Gemma 3 12B Instruct → 12B | ~8 GB | 128,000 | ollama run gemma3:12b | |
| Gemma 3 27B Instruct → 27B | ~17.1 GB | 128,000 | ollama run gemma3:27b |
Memory is quantized weights plus overhead at Q4_K_M, from the same engine as the GPU & VRAM checker.
How to run Gemma 3 locally
Install Ollama, then pull the tag.
ollama run gemma3:1bPick a size above for its own VRAM figure, speed estimate and install command.
Licence
Weights are downloadable and commercial use is permitted, subject to the licence’s acceptable-use terms.
Applies to: Gemma 3 1B Instruct, Gemma 3 4B Instruct, Gemma 3 12B Instruct, Gemma 3 27B InstructRecommended GPU
The cheapest catalogued GPU that runs Gemma 3 locally (min 1 GB VRAM) is the Intel Arc B570 (10 GB).
Can I run Gemma 3 on my GPU?
- Gemma 3 on AMD Radeon RX 7800 XT
- Gemma 3 on AMD Radeon RX 7900 XTX
- Gemma 3 on AMD Radeon RX 9060 XT 16GB
- Gemma 3 on AMD Radeon RX 9070
- Gemma 3 on AMD Radeon RX 9070 XT
- Gemma 3 on Apple M1
- Gemma 3 on Apple M1 Pro
- Gemma 3 on Apple M2 Pro
- Gemma 3 on Apple M3 Pro
- Gemma 3 on Apple M4
- Gemma 3 on Apple M5
- Gemma 3 on Intel Arc B570
- Gemma 3 on Intel Arc B580
- Gemma 3 on NVIDIA GeForce RTX 3060 (12GB)
- Gemma 3 on NVIDIA GeForce RTX 3080 (10GB)
- Gemma 3 on NVIDIA GeForce RTX 3090
- Gemma 3 on NVIDIA GeForce RTX 4060 Ti 16GB
- Gemma 3 on NVIDIA GeForce RTX 4070
- Gemma 3 on NVIDIA GeForce RTX 4070 Super
- Gemma 3 on NVIDIA GeForce RTX 4070 Ti
- Gemma 3 on NVIDIA GeForce RTX 4070 Ti Super
- Gemma 3 on NVIDIA GeForce RTX 4080
- Gemma 3 on NVIDIA GeForce RTX 4080 Super
- Gemma 3 on NVIDIA GeForce RTX 4090
- Gemma 3 on NVIDIA GeForce RTX 5060 Ti 16GB
- Gemma 3 on NVIDIA GeForce RTX 5070
- Gemma 3 on NVIDIA GeForce RTX 5070 Ti
- Gemma 3 on NVIDIA GeForce RTX 5080
- Gemma 3 on NVIDIA GeForce RTX 5090
Gemma 3 — frequently asked questions
How much VRAM does Gemma 3 need?
Gemma 3 needs about 1 GB VRAM at Q4_K_M quantization for its smallest variant. Variants: Gemma 3 1B Instruct (1 GB, Q4_K_M); Gemma 3 4B Instruct (3 GB, Q4_K_M); Gemma 3 12B Instruct (8 GB, Q4_K_M); Gemma 3 27B Instruct (17 GB, Q4_K_M). On Apple Silicon, unified memory counts toward this requirement.
Can I run Gemma 3 on an RTX 4090 (24 GB)?
Yes — Gemma 3 runs on an RTX 4090 (24 GB) and other 24 GB cards such as the RTX 3090. Smaller variants also fit comfortably on 8–16 GB GPUs at Q4_K_M.
What quantization should I use for Gemma 3?
Q4_K_M is the best balance of quality and VRAM for Gemma 3 in most cases. Choose Q8_0 for near-lossless quality if you have spare VRAM, or smaller quants (Q3/Q2) only when memory is tight.
How do I run Gemma 3 with Ollama?
Install Ollama, then run: ollama run gemma3:1b. This downloads Gemma 3 and starts a local, OpenAI-compatible endpoint — no internet connection is needed after the initial download.