Skip to content

Gemma 4 26B A4B on RTX 3090: VRAM, speed and fit

Gemma 4 26B A4B at 6-bit runs on the RTX 3090 with 184K of context at about 153 tokens a second.

Text generation · Chat · Ollama / LM Studio

Runs comfortably · on RTX 3090 · 6-bit (Q6_K) · Ollama / LM Studio · Chat

GPU memory
22.7 GB
of 24.0 GB
Context
184K
auto
Speed
153 tok/s
102 when full
First reply
0.7 s
2K prompt

Where the memory goes

Model
19.7 GB
Context
2.0 GB
Overhead
1.0 GB
OS
0.0 GB
Free
1.3 GB
Memory
22.7 GB of 24.0 GB
Context
184K tokens · auto
Speed
153 tok/s · 102 when full
First reply
0.7 s for a 2K prompt
Running cost
$6.35/mo at home · $33/mo on a cloud RTX A5000

Validation: Validated

What Gemma 4 26B A4B on RTX 3090 was checked against
CheckTargetResultStatus
Fits in GPU memory24.0 GB usable22.7 GB neededPass
Headroom5% free1.3 GB free (5%)Pass
Context for chat16K or more184KPass
Writing speed15 tok/s or more153 tok/sPass
First replyunder 3 s, 2K prompt0.7 sPass

Targets: a reply you can read along with, a first reply the use case tolerates, and 5% of memory left free. Calculated from published specs, not benchmarked.

The setup

Model
Gemma 4 26B A4B
Parameters
25.8B, 3.8B active
Precision
6-bit (Q6_K)
Context
184K, auto
Chats at once
1
Use
Chat
Runs in
Ollama / LM Studio
Hardware
RTX 3090
GPU can use
24.0 GB of 24.0 GB

Running cost

At home
$6.35/mo
290 W at the wall
Renting instead
$33/mo
RTX A5000 at $0.27/h

4 h a day, power at $0.18 per kWh, cloud at RunPod on-demand prices as of Nov 21, 2026.

How to run it

Ollama
ollama run gemma4:26b
llama.cpp
llama-server -hf unsloth/gemma-4-26B-A4B-it-GGUF
vLLM
vllm serve google/gemma-4-26B-A4B-it
MLX (Mac)
mlx_lm.chat --model mlx-community/gemma-4-26b-a4b-it-4bit

Gemma 4 26B A4B on similar machines

All machines: Gemma 4 26B A4B VRAM requirements. Everything that runs here: what can the RTX 3090 run?

Other models that run on RTX 3090

Questions

Can Gemma 4 26B A4B run on RTX 3090?

Gemma 4 26B A4B at 6-bit runs on the RTX 3090 with 184K of context at about 153 tokens a second.

Report VR-1LBVY6A, calculated Nov 22, 2026 from published specs, not benchmarked.