Skip to content

Diffusiongemma 26B A4B IT on 2× RTX 3090: VRAM, speed and fit

Diffusiongemma 26B A4B IT at 8-bit runs on the 2× RTX 3090 with 256K of context at about 121 tokens a second.

Text generation · Chat · Ollama / LM Studio

Runs comfortably · on 2× RTX 3090 · 8-bit (Q8_0) · Ollama / LM Studio · Chat

GPU memory
32.5 GB
of 48.0 GB
Context
256K
auto
Speed
121 tok/s
60 when full
First reply
0.4 s
2K prompt

Where the memory goes

Model
25.6 GB
Context
5.2 GB
Overhead
1.8 GB
OS
0.0 GB
Free
15.5 GB
Memory
32.5 GB of 48.0 GB
Context
256K tokens · auto
Speed
121 tok/s · 60 when full
First reply
0.4 s for a 2K prompt
Running cost
$11/mo at home · $61/mo on a cloud 2× RTX A4500

Validation: Validated

What Diffusiongemma 26B A4B IT on 2× RTX 3090 was checked against
CheckTargetResultStatus
Fits in GPU memory48.0 GB usable32.5 GB neededPass
Headroom5% free15.5 GB free (32%)Pass
Context for chat16K or more256KPass
Writing speed15 tok/s or more121 tok/sPass
First replyunder 3 s, 2K prompt0.4 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
Diffusiongemma 26B A4B IT
Parameters
25.8B, 4B active
Precision
8-bit (Q8_0)
Context
256K, auto
Chats at once
1
Use
Chat
Runs in
Ollama / LM Studio
Hardware
2× RTX 3090
GPU can use
48.0 GB of 48.0 GB

Running cost

At home
$11/mo
500 W at the wall
Renting instead
$61/mo
2× RTX A4500 at $0.50/h

4 h a day, power at $0.18 per kWh, cloud at RunPod on-demand prices as of Nov 22, 2026.

How to run it

llama.cpp
llama-server -hf unsloth/diffusiongemma-26B-A4B-it-GGUF
vLLM
vllm serve google/diffusiongemma-26B-A4B-it
MLX (Mac)
mlx_lm.chat --model mlx-community/diffusiongemma-26B-A4B-it-4bit

Diffusiongemma 26B A4B IT on similar machines

All machines: Diffusiongemma 26B A4B IT VRAM requirements. Everything that runs here: what can the 2× RTX 3090 run?

Other models that run on 2× RTX 3090

Questions

Can Diffusiongemma 26B A4B IT run on 2× RTX 3090?

Diffusiongemma 26B A4B IT at 8-bit runs on the 2× RTX 3090 with 256K of context at about 121 tokens a second.

Report VR-05EPELR, calculated Nov 22, 2026 from published specs, not benchmarked.