Gemma 4 26B-A4B — VRAM, speed & local setup
Written by Jakub Rusinowski · Last updated
Gemma 4's Mixture-of-Experts variant: 26B total parameters but only 4B active per forward pass. This gives 12B-class quality at 4B-class inference speed. Supports text, image, and audio. Needs ~16–18 GB at Q4_K_M — fits in an RTX 3090 or 4090. Excellent for developers who need a step above the E4B without occupying the full GPU.
Gemma 4 26B-A4B needs about 16 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 | 9.8 GB | ~222 tok/s | Fits | |
| Q3_K_M 3.41 bpw | 12.4 GB | ~199 tok/s | Fits | |
| Q4_K_M 4.83 bpw | 17 GB | ~168 tok/s | Fits | |
| Q5_K_M 5.67 bpw | 19.7 GB | ~153 tok/s | Fits | |
| Q6_K 6.56 bpw | 22.6 GB | ~141 tok/s | Tight | |
| Q8_0 8.50 bpw | 28.9 GB | ~18 tok/s | Offload | |
| F16 16.00 bpw | 53.3 GB | ~11 tok/s | Offload |
Black marker = usable memory on the NVIDIA RTX 4090 (24 GB). Estimates from the memory-bandwidth roofline on the methodology page. Gemma 4 26B-A4B VRAM calculator →
Get Gemma 4 26B-A4B running
The cheapest catalogued GPU that runs Gemma 4 26B-A4B is the AMD Radeon RX 7900 XT (20 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 4 26B-A4B
Install Ollama, then run:
ollama run gemma4:26b-a4bSpecifications
Verified — Checked against the primary source — the model card or the vendor spec page — and corroborated by a second independent source.
- Parameters
- 26 Billion (4B active per token)
- Context window
- 256,000
- Architecture
- Mixture-of-Experts + Multimodal Encoder (image + audio)
- Provider
- Licence
- Apache 2.0
- Practical deployment
- Single consumer GPU
- Specified at
- Q4_K_M
- System RAM
- 32 GB
- Record updated
- 2026-03-31
Commercial use permitted. No usage restrictions beyond attribution.
Quality and use cases
Scores as published by the model’s authors or an independent evaluator — quality, not throughput, and not measured by us.
| Benchmark | Score | Provenance |
|---|---|---|
| MMLU | 87.5 / 100 % | reported |
| HumanEval | 83.7 / 100 % | reported |
| MATH | 74.2 / 100 % | reported |
Can I run Gemma 4 26B-A4B on my GPU?
- Gemma 4 26B-A4B on AMD Radeon RX 7800 XT
- Gemma 4 26B-A4B on AMD Radeon RX 7900 XT
- Gemma 4 26B-A4B on AMD Radeon RX 7900 XTX
- Gemma 4 26B-A4B on AMD Radeon RX 9060 XT 16GB
- Gemma 4 26B-A4B on AMD Radeon RX 9060 XT 8GB
- Gemma 4 26B-A4B on AMD Radeon RX 9070
- Gemma 4 26B-A4B on AMD Radeon RX 9070 XT
- Gemma 4 26B-A4B on Apple M1
- Gemma 4 26B-A4B on Apple M1 Pro
- Gemma 4 26B-A4B on Apple M2
- Gemma 4 26B-A4B on Apple M2 Pro
- Gemma 4 26B-A4B on Apple M3
- Gemma 4 26B-A4B on Apple M3 Pro
- Gemma 4 26B-A4B on Apple M4
- Gemma 4 26B-A4B on Apple M4 Pro
- Gemma 4 26B-A4B on Apple M5
- Gemma 4 26B-A4B on Intel Arc B570
- Gemma 4 26B-A4B on Intel Arc B580
- Gemma 4 26B-A4B on NVIDIA GeForce RTX 3060 (12GB)
- Gemma 4 26B-A4B on NVIDIA GeForce RTX 3070
- Gemma 4 26B-A4B on NVIDIA GeForce RTX 3070 Ti
- Gemma 4 26B-A4B on NVIDIA GeForce RTX 3080 (10GB)
- Gemma 4 26B-A4B on NVIDIA GeForce RTX 3090
- Gemma 4 26B-A4B on NVIDIA GeForce RTX 4060
- Gemma 4 26B-A4B on NVIDIA GeForce RTX 4060 Ti 16GB
- Gemma 4 26B-A4B on NVIDIA GeForce RTX 4070
- Gemma 4 26B-A4B on NVIDIA GeForce RTX 4070 Super
- Gemma 4 26B-A4B on NVIDIA GeForce RTX 4070 Ti
- Gemma 4 26B-A4B on NVIDIA GeForce RTX 4070 Ti Super
- Gemma 4 26B-A4B on NVIDIA GeForce RTX 4080
- Gemma 4 26B-A4B on NVIDIA GeForce RTX 4080 Super
- Gemma 4 26B-A4B on NVIDIA GeForce RTX 4090
- Gemma 4 26B-A4B on NVIDIA GeForce RTX 5060
- Gemma 4 26B-A4B on NVIDIA GeForce RTX 5060 Ti 16GB
- Gemma 4 26B-A4B on NVIDIA GeForce RTX 5060 Ti 8GB
- Gemma 4 26B-A4B on NVIDIA GeForce RTX 5070
- Gemma 4 26B-A4B on NVIDIA GeForce RTX 5070 Ti
- Gemma 4 26B-A4B on NVIDIA GeForce RTX 5080
- Gemma 4 26B-A4B on NVIDIA GeForce RTX 5090
Other Gemma 4 sizes
Gemma 4 26B-A4B — frequently asked questions
How much VRAM does Gemma 4 26B-A4B need?
About 16 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 4 26B-A4B run on an RTX 4090 (24 GB)?
Yes. Gemma 4 26B-A4B needs about 16 GB at Q4_K_M, inside a 24 GB card, at an estimated 168 tokens/sec.
How do I run Gemma 4 26B-A4B locally?
Install Ollama and run `ollama run gemma4:26b-a4b`. That pulls the weights and starts a local OpenAI-compatible endpoint; after the download nothing leaves the machine.
What other sizes does Gemma 4 come in?
Gemma 4 E2B (4 GB), Gemma 4 E4B (6 GB), Gemma 4 26B-A4B (16 GB), Gemma 4 31B (20 GB), Gemma 4 12B (Unified) (8 GB). Every size shares the family's training and licence; the larger ones score higher and need proportionally more memory.