Gemma 3n E2B — VRAM, speed & local setup
Written by Jakub Rusinowski · Last updated
The smallest Gemma 3n variant. Only 2 GB RAM required, designed to run on any modern smartphone. Supports vision (images) as input.
Gemma 3n E2B needs about 4 GB of VRAM at Q4_K_M — quantized weights plus framework overhead, before any KV cache. On Apple Silicon that figure comes out of unified memory.
VRAM and speed by quantization
Quoted against NVIDIA RTX 4090 (24 GB). Weights plus overhead only: this model's architecture is unpublished, so no KV cache is added.
| Quant | Memory | VRAM | Speed (est.) | Fit |
|---|---|---|---|---|
| Q2_K 2.63 bpw | 2.6 GB | ~267 tok/s | Fits | |
| Q3_K_M 3.41 bpw | 3.1 GB | ~251 tok/s | Fits | |
| Q4_K_M 4.83 bpw | 4.1 GB | ~225 tok/s | Fits | |
| Q5_K_M 5.67 bpw | 4.7 GB | ~213 tok/s | Fits | |
| Q6_K 6.56 bpw | 5.3 GB | ~201 tok/s | Fits | |
| Q8_0 8.50 bpw | 6.6 GB | ~179 tok/s | Fits | |
| F16 16.00 bpw | 11.7 GB | ~126 tok/s | Fits |
Black marker = usable memory on the NVIDIA RTX 4090 (24 GB). Estimates from the memory-bandwidth roofline on the methodology page. Gemma 3n E2B VRAM calculator →
Get Gemma 3n E2B running
The cheapest catalogued GPU that runs Gemma 3n E2B is the Intel Arc B570 (10 GB).
or compare on Vast.ai from $0.35/hr (typical low · varies)
As an Amazon Associate we earn from qualifying purchases. Cloud GPU links are referral links — we may earn a commission at no extra cost to you.
How to run Gemma 3n E2B
Install Ollama, then run:
ollama run gemma3n:e2bSpecifications
Curated — A hand-written entry from before this catalogue recorded its sources. The figures are long-standing but their provenance is not on file.
- Parameters
- 2 Billion effective
- Context window
- 32,768
- Architecture
- MatFormer, multimodal
- Provider
- Google DeepMind
- Licence
- Gemma ToS
- Practical deployment
- Edge device (phone, SBC)
- Specified at
- Q4_K_M
- System RAM
- 4 GB
- Record updated
- 2025-04-01
Weights are downloadable and commercial use is permitted, subject to the licence’s acceptable-use terms.
Quality and use cases
Scores as published by the model’s authors or an independent evaluator — quality, not throughput, and not measured by us.
Can I run Gemma 3n E2B on my GPU?
- Gemma 3n E2B on AMD Radeon RX 9060 XT 8GB
- Gemma 3n E2B on Intel Arc B570
- Gemma 3n E2B on NVIDIA GeForce RTX 3070
- Gemma 3n E2B on NVIDIA GeForce RTX 3070 Ti
- Gemma 3n E2B on NVIDIA GeForce RTX 3080 (10GB)
- Gemma 3n E2B on NVIDIA GeForce RTX 4060
- Gemma 3n E2B on NVIDIA GeForce RTX 5060
- Gemma 3n E2B on NVIDIA GeForce RTX 5060 Ti 8GB
Other Gemma 3n sizes
Gemma 3n E2B — frequently asked questions
How much VRAM does Gemma 3n E2B need?
About 4 GB at Q4_K_M — quantized weights plus framework overhead, before any KV cache. The cache grows with context length and is added on top; the table above folds it in. Apple Silicon counts unified memory toward the same figure.
Does Gemma 3n E2B run on an RTX 4090 (24 GB)?
Yes. Gemma 3n E2B needs about 4 GB at Q4_K_M, inside a 24 GB card, at an estimated 225 tokens/sec.
How do I run Gemma 3n E2B locally?
Install Ollama and run `ollama run gemma3n:e2b`. That pulls the weights and starts a local OpenAI-compatible endpoint; after the download nothing leaves the machine.
What other sizes does Gemma 3n come in?
Gemma 3n E2B (4 GB), Gemma 3n E4B (6 GB). Every size shares the family's training and licence; the larger ones score higher and need proportionally more memory.