Gemma 2 9B IT — VRAM, speed & local setup
Written by Jakub Rusinowski · Last updated
Punchy and creative. Excellent for creative writing and brainstorming.
Gemma 2 9B IT needs about 6 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). Includes the KV cache at 8K context, so it reads higher than the headline figure.
| Quant | Memory | VRAM | Speed (est.) | Fit |
|---|---|---|---|---|
| Q2_K 2.63 bpw | 6 GB | ~125 tok/s | Fits | |
| Q3_K_M 3.41 bpw | 6.8 GB | ~109 tok/s | Fits | |
| Q4_K_M 4.83 bpw | 8.4 GB | ~89 tok/s | Fits | |
| Q5_K_M 5.67 bpw | 9.4 GB | ~80 tok/s | Fits | |
| Q6_K 6.56 bpw | 10.4 GB | ~72 tok/s | Fits | |
| Q8_0 8.50 bpw | 12.6 GB | ~60 tok/s | Fits | |
| F16 16.00 bpw | 21 GB | ~36 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 2 9B IT VRAM calculator →
Get Gemma 2 9B IT running
The cheapest catalogued GPU that runs Gemma 2 9B IT 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 2 9B IT
Install Ollama, then run:
ollama run gemma2Download Gemma 2 9B IT — GGUF Quantizations
Pick a quantization and open it in LM Studio, Ollama, or Jan, or download the raw .gguf file directly. Quant list and sizes resolved from Hugging Face.
Gemma 2 9B IT — GGUF quants · bartowski/gemma-2-9b-it-GGUF
| Quant | Size | Download (.gguf) |
|---|---|---|
| Q3_K_M | 3.84 GB (est.) | gemma-2-9b-it-Q3_K_M.gguf |
| Q4_K_M | 5.43 GB (est.) | gemma-2-9b-it-Q4_K_M.gguf |
| Q5_K_M | 6.38 GB (est.) | gemma-2-9b-it-Q5_K_M.gguf |
| Q6_K | 7.38 GB (est.) | gemma-2-9b-it-Q6_K.gguf |
| Q8_0 | 9.56 GB (est.) | gemma-2-9b-it-Q8_0.gguf |
Download in LM Studio: lms get bartowski/gemma-2-9b-it-GGUF
Want this model on your phone? You can run it on your desktop with LM Studio and chat from your iPhone or iPad over an encrypted link — see Run LM Studio Models on Your Phone (LM Link).
Specifications
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
- 9 Billion
- Context window
- 8,192
- Architecture
- Dense Transformer
- Provider
- Licence
- Gemma Terms
- Practical deployment
- Laptop
- Specified at
- Q4_K_M
- System RAM
- 16 GB
- Record updated
- 2024-06-27
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 2 9B IT on my GPU?
- Gemma 2 Family on AMD Radeon RX 9060 XT 8GB
- Gemma 2 Family on Intel Arc B570
- Gemma 2 Family on Intel Arc B580
- Gemma 2 Family on NVIDIA GeForce RTX 3060 (12GB)
- Gemma 2 Family on NVIDIA GeForce RTX 3070
- Gemma 2 Family on NVIDIA GeForce RTX 3070 Ti
- Gemma 2 Family on NVIDIA GeForce RTX 3080 (10GB)
- Gemma 2 Family on NVIDIA GeForce RTX 4060
- Gemma 2 Family on NVIDIA GeForce RTX 4070
- Gemma 2 Family on NVIDIA GeForce RTX 4070 Super
- Gemma 2 Family on NVIDIA GeForce RTX 4070 Ti
- Gemma 2 Family on NVIDIA GeForce RTX 5060
- Gemma 2 Family on NVIDIA GeForce RTX 5060 Ti 8GB
- Gemma 2 Family on NVIDIA GeForce RTX 5070
Gemma 2 9B IT — frequently asked questions
How much VRAM does Gemma 2 9B IT need?
About 6 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 2 9B IT run on an RTX 4090 (24 GB)?
Yes. Gemma 2 9B IT needs about 6 GB at Q4_K_M, inside a 24 GB card, at an estimated 89 tokens/sec.
How do I run Gemma 2 9B IT locally?
Install Ollama and run `ollama run gemma2`. That pulls the weights and starts a local OpenAI-compatible endpoint; after the download nothing leaves the machine.