Gemma 4 26B-A4B VRAM requirements

Can I run Gemma 4 26B-A4B on…

• RTX 4090: Yes, up to Q6_K
• RTX 3090: Yes, up to Q6_K
• RTX 5090: Yes, up to Q8_0
• RTX 3060 12GB: --n-cpu-moe 12
• RTX 4060 Ti 16GB: Only Q3_K_M
• RTX 5060 Ti 16GB: Only Q3_K_M
• RTX 4070 12GB: --n-cpu-moe 12
• RTX 5080 16GB: Only Q3_K_M
• RTX 4080 Super 16GB: Only Q3_K_M
• RX 7900 XTX: Yes, up to Q6_K

Running Gemma 4 26B-A4B with experts in system RAM

llama.cpp's --n-cpu-moe N keeps the experts of the first N layers in RAM. For the measured UD-Q4_K_M GGUF with 32K tokens of context, the smallest N that fits each card, and the speed with 89.6 GB/s of system RAM. Plan the offload for another GPU. The VRAM here can sit below the 17.5 GB of the worked example because it counts the file's measured bytes, keeps the input embedding (token_embd) in system RAM as llama.cpp does and adds a flat 1 GB of buffers, where the estimate above adds 0.5 GB plus 10% to the preset's weights.

In system RAM counts the offloaded experts plus the input embedding (token_embd), which stays in RAM even with --n-cpu-moe 0.

How much VRAM does Gemma 4 26B-A4B need?

Gemma 4 26B-A4B needs 17.0 GB at Q4_K_M, 29.1 GB at Q8_0 and 53.9 GB at FP16 with 8,192 tokens of context and one request. Each total below is the weights plus the FP16 KV cache and the runtime overhead (0.5 GB plus 10%), and each row opens the calculator with that setting. What the GGUF names mean.

Fine-tuning Gemma 4 26B-A4B? Gemma 4 26B-A4B VRAM for LoRA, QLoRA and full training.

Gemma 4 26B-A4B GGUF files on Hugging Face

The 8 GGUF files llama.cpp picks for Gemma 4 26B-A4B with -hf <repo>:<QUANT>, with the bytes the Hugging Face file list reported on 2026-09-29. Each total is that file plus the FP16 KV cache and 0.5 GB + 10%, one request. "vs estimate" compares the file with the bits-per-weight size the table above uses.

GB here is GiB (1024³ bytes), as everywhere on this site; Hugging Face shows the Q4_K_M file as 16.9 GB, in 10⁹ bytes.

Which Gemma 4 26B-A4B GGUF fits a 16, 24, 32, 48 or 80 GB GPU or a Mac

From the file sizes above: the largest file that fits each machine with an 8K context and 0.5 GB left free, the longest context it then has room for, and the longest context for Q4_K_M (15.8 GB). One request, FP16 KV cache, -np 1. Where nothing fits, the --n-cpu-moe table above keeps experts in system RAM.

Run Gemma 4 26B-A4B with llama.cpp or Ollama

llama-server -hf unsloth/gemma-4-26B-A4B-it-GGUF:Q4_K_M -c 252928 -np 1

gemma-4-26B-A4B-it-UD-Q4_K_M.gguf, 16.9 GB, from unsloth/gemma-4-26B-A4B-it-GGUF (checked 2026-09-29). On a 24 GB RTX 3090 / 4090, Q4_K_M with -c 252928 takes 23.5 GB and leaves 521 MB. The file is 1.24 GB over the Q4_K_M estimate, so -c counts the file. -np 1: one slot, so one 1,024-token window.

ollama run gemma4:26b

Ollama library tag gemma4:26b; its quant is Ollama's choice. Same context: /set parameter num_ctx 252928.

Gemma 4 26B-A4B at 8K, 32K, 128K and 256K (full) tokens of context

5 of its 30 layers use full attention and 25 keep a sliding window of 1,024 tokens. Each extra token of context adds 20 KB of FP16 cache per request once the sliding windows are full. The last column is the smallest setup here, one card or a group, that holds it at Q4_K_M. How the KV cache works.

How fast Gemma 4 26B-A4B writes

Tokens per second for one request with 8,192 tokens of context, estimated from memory bandwidth and the parameters read per token. A dash means it does not fit on one card. Try other settings in the speed calculator.

Longest context on one GPU

How many tokens of context Gemma 4 26B-A4B fits on a single card with one request, an FP16 KV cache and 0.5 GB left free. "Full" means the model's whole context window fits.

Model details

Parameters

25.8B (25,805,936,206)

Experts

128 routed experts, all loaded

Active per token

4B

Layers

5 of its 30 layers use full attention and 25 keep a sliding window of 1,024 tokens

Attention cache

2 KV heads × 512 (values come from the key projection but are cached separately)

Sliding-window layers

8 KV heads × 256

Context length

262,144 tokens

Published weights

48.1 GB (BF16)

On Hugging Face

google/gemma-4-26B-A4B-it

Reported benchmarks

GPQA Diamond 82.3 · MMLU-Pro 82.6 · LiveCodeBench v6 77.1. Vendor-reported: from the maker’s own model card, not rerun here; read 2026-09-30.

Gemma 4 26B-A4B next to similar-sized models

The three models here closest to it in parameter count, at Q4_K_M with 8,192 tokens of context.