edwixxDiffusionGemmaDiffusionGemmaForBlockDiffusion

Diffusiongemma 26B A4B IT HERETIC Uncensored — Hardware Requirements & GPU Compatibility

Chat

Diffusiongemma 26B A4B IT HERETIC Uncensored is a 25.8B-parameter open language model from edwixx in the DiffusionGemma family. It supports a context window of up to 262,144 tokens. At Q4_K_M it needs about 16.14 GB of VRAM — see which GPUs and Macs can run it below.

25.8B params · MoE4.4B active262K contextApache 2.0Jun 202613 downloads3 likes3.9K quant downloads

Based on Diffusiongemma 26B A4B IT

Your hardware

Verdicts below are for the reference RTX 4090. Set yours to personalise the page.

Detection reads your browser's GPU name with no data sent anywhere.

S

Runs great on a 24 GB GPU

Runs from 12 GB (Q2_K · 16K)Best on 24 GB (Q4_K_M · ~171 tok/s · 128K)Macs from 16 GB (Q2_K · ~50 tok/s)

Apache 2.0

Reference GPU · RTX 4090 24 GB

Estimates, ±20% · how we estimate

Run Diffusiongemma 26B A4B IT HERETIC Uncensored

One recommended way. Every other method, OS and quant is in the step-by-step guide.Each way opens a step-by-step guide for your OS and quant.

Recommended

Ollama

one command · needs Ollama installed

$ollama run hf.co/yabaimimi/edwixx__diffusiongemma-26B-A4B-it-HERETIC-Uncensored-GGUF:Q4_K_M

Q4_K_M fits fully on the RTX 4090 · 7.5 GB free · 16K context

Download: edwixx__diffusiongemma-26B-A4B-it-HERETIC-Uncensored-Q4_K_M.gguf16.81 GByabaimimi

install, OS, other quants

Ways to runOther ways

each opens step-by-step

Run in cloud

Your own private copy · Q4_K_M · 24 GB GPU

The easiest way to try this model
Running in a few minutes, easy setup
Chat right here or connect any app

$5+ bonus for new users

How Much VRAM Does Diffusiongemma 26B A4B IT HERETIC Uncensored Need?

Select a quantization level; the listed hardware adheres to it.

Quantization	VRAM	+32K	+128K	File	Quality	On RTX 4090 (reference)
	11.6 GB	12.5 GB	15.8 GB	10.6 GB	2-bit · K-quant — 2-bit quantization with K-quant improvements	Fits: 256K context · ~214 tok/s
	13.2 GB	14.1 GB	17.5 GB	13.3 GB	3-bit medium — 3-bit medium quantization	Fits: 256K context · ~196 tok/s
	16.1 GB	17 GB	20.4 GB	16.8 GB	4-bit medium — 4-bit medium quantization — most popular sweet spot	Fits: 128K context · ~171 tok/s
	19.1 GB	20 GB	23.3 GB	19.1 GB	5-bit medium — 5-bit medium quantization — good quality/size tradeoff	Fits: 128K context · ~151 tok/s
	21.9 GB	22.9 GB	26.2 GB	22.7 GB	6-bit, very good — 6-bit quantization, very good quality	Fits: 64K context · ~136 tok/s
	26.5 GB	27.4 GB	30.7 GB	26.9 GB	8-bit, near-lossless — 8-bit quantization, near-lossless	Does not fit: needs 26.8 GB
	52.3 GB	53.2 GB	56.5 GB	50.5 GB	16-bit · bfloat — Brain floating point 16 — preferred for training	Does not fit: needs 52.6 GB

est.calculated, no published file yet

Selected Q4_K_M

4-bit medium quantization

16.5GBat 16K context

Weights: 15.5 GB
KV cache: 0.7 GB
Overhead: 0.3 GB

Fits RTX 4090 (reference) · 7.5 GB free · ~171 tok/s

≥50 tok/s needs ≥231 GB/s · RTX 5090 ~249, RTX 3090 Ti ~171

Which Hardware Can Run Diffusiongemma 26B A4B IT HERETIC Uncensored?

Q4_K_M · 16.1 GB

Diffusiongemma 26B A4B IT HERETIC Uncensored (Q4_K_M) requires 16.1 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 21+ GB is recommended. Using the full 262K context window can add up to 8.6 GB, bringing total usage to 24.8 GB. 8 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti. 122 devices with unified memory can run Diffusiongemma 26B A4B IT HERETIC Uncensored, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), Mac Mini M5 Pro (24 GB).

By VRAM tier · typical GPUs and Macs at Q4_K_M

Typical hardwareQuantVRAMtok/sContext that fits

8 GB · RTX 4060Doesn't fit — smallest quant needs 12 GB8 GB tier →

12 GB · RTX 3060 12GBQ2_K · 12 GB · ~99 tok/s · 16K12~9916K12 GB tier →

16 GB · RTX 5060 Ti 16GBQ3_K_M · 13.6 GB · ~106 tok/s · 64K13.6~10664K16 GB tier →

24 GB · RTX 4090referenceQ4_K_M · 16.5 GB · ~171 tok/s · 128K16.5~171128K24 GB tier →

Mac 32 GB · Mac Mini M4Q4_K_M · 16.5 GB · ~26 tok/s · 256K16.5~26256KMacs →

Mac 64 GB · Mac Studio M4 MaxQ4_K_M · 16.5 GB · ~76 tok/s · 256K16.5~76256KMacs →

Alternatives to Diffusiongemma 26B A4B IT HERETIC Uncensored

Similar models

6tok/s on RTX 4090

Model	Score	VRAM	tok/s	Why it is similar
Qwen3.8 27BAlibaba	122.0	17.9 GB	~38	same size
Qwen3.6 27BAlibaba	111.2	17.9 GB	~38	same size
Qwen3.5 27BAlibaba	103.8	17.9 GB	~38	same size
LFM2.5 24B A2BLiquid AI	—	15 GB	~221	same size
Gemma 4 26B A4BGoogle	—	16.5 GB	~171	same size
Devstral Small 2 24B Instruct 2512Mistral AI	—	18.1 GB	~43	same size

Specifications

Publisher: edwixx
Family: DiffusionGemma
Parameters: 25.8B
Active parameters: 4.4B per token
Architecture: DiffusionGemmaForBlockDiffusion
Context Length: 262,144 tokens
Vocabulary Size: 262,144
Release Date: 2026-06-11
License: Apache 2.0

Resources

Hugging Face model card (opens in a new tab)

Frequently Asked Questions

How much VRAM does Diffusiongemma 26B A4B IT HERETIC Uncensored need?

Diffusiongemma 26B A4B IT HERETIC Uncensored requires 16.1 GB of VRAM at Q4_K_M, or 52.3 GB at BF16. Full 262K context adds up to 8.6 GB (24.8 GB total).

VRAM = Weights + KV Cache + Overhead

Weights = 25.8B × 4.8 bits ÷ 8 = 15.5 GB

KV Cache + Overhead ≈ 0.6 GB (at 2K context + ~0.3 GB framework)

Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.

KV Cache + Overhead ≈ 9.3 GB (at full 262K context)

VRAM usage by quantization

Q4_K_M

16.1 GB

Q4_K_M + full context

24.8 GB

Learn more about VRAM estimation →

Can NVIDIA GeForce RTX 4090 run Diffusiongemma 26B A4B IT HERETIC Uncensored?

Yes, at Q6_K (21.9 GB) or lower. Higher quantizations like Q8_0 (26.5 GB) exceed the NVIDIA GeForce RTX 4090's 24 GB.

What's the best quantization for Diffusiongemma 26B A4B IT HERETIC Uncensored?

For Diffusiongemma 26B A4B IT HERETIC Uncensored, Q4_K_M (16.1 GB) offers the best balance of quality and VRAM usage. Q5_K_M (19.1 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 11.6 GB.

VRAM requirement by quantization

Q2_K

11.6 GB~75%

Q4_K_M ★

16.1 GB~89%

Q5_K_M

19.1 GB~92%

Q6_K

21.9 GB~95%

Q8_0

26.5 GB~99%

BF16

52.3 GB~100%

★ Recommended — best balance of quality and VRAM usage.

Learn more about quantization →

Can I run Diffusiongemma 26B A4B IT HERETIC Uncensored on a Mac?

Yes — Mac Mini M2 Pro (16 GB) and 133 other Macs can run Diffusiongemma 26B A4B IT HERETIC Uncensored, the smallest at Q2_K. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 16.1 GB of usable unified memory (RAM minus macOS overhead), from Mac Mini M5 Pro (24 GB).

Can I run Diffusiongemma 26B A4B IT HERETIC Uncensored locally?

Yes — Diffusiongemma 26B A4B IT HERETIC Uncensored can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.

How fast is Diffusiongemma 26B A4B IT HERETIC Uncensored?

At Q4_K_M, Diffusiongemma 26B A4B IT HERETIC Uncensored can reach ~152 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~171 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.

tok/s = 1000 ÷ (active GB ÷ (bandwidth GB/s × efficiency) × 1000 + layers × routing ms)

Mixture-of-Experts: just the active experts get read for each token, along with a fixed routing cost per layer.

Example: NVIDIA B200 → 2.8 GB active ÷ (8000 × 0.65) = 0.53 ms, plus 30 layers × 0.055 ms = 1.65 ms, so 1000 ÷ 2.18 ms = ~459 tok/s

Estimated speed at Q4_K_M (16.1 GB)

NVIDIA B200

~459 tok/s

NVIDIA GeForce RTX 4090

~171 tok/s

NVIDIA B300

~459 tok/s

NVIDIA GH200 Grace Hopper Superchip

~398 tok/s

Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.

Learn more about tok/s estimation →

What's the download size of Diffusiongemma 26B A4B IT HERETIC Uncensored?

At Q4_K_M, the download is about 15.49 GB. The full-precision BF16 version is 51.65 GB. The smallest option (Q2_K) is 10.98 GB.

Which GPUs can run Diffusiongemma 26B A4B IT HERETIC Uncensored?

8 consumer GPUs can run Diffusiongemma 26B A4B IT HERETIC Uncensored at Q4_K_M. Top options include NVIDIA GeForce RTX 5090, AMD Radeon RX 7900 XT, AMD Radeon RX 7900 XTX. 1 GPU has plenty of headroom for comfortable inference.

Which devices can run Diffusiongemma 26B A4B IT HERETIC Uncensored?

122 devices with unified memory can run Diffusiongemma 26B A4B IT HERETIC Uncensored at Q4_K_M (16.1 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.

Run Diffusiongemma 26B A4B IT HERETIC Uncensored

16.5 GB VRAM at 16K context · fits the RTX 4090 (7.5 GB free)

Works on macOS, Windows and Linux.

Install Ollama (Linux)
$curl -fsSL https://ollama.com/install.sh | sh
Install Ollama (macOS / Windows)
Download and run the app; it installs the ollama command.
Download Ollama ↗
Run Q4_K_M
$ollama run hf.co/yabaimimi/edwixx__diffusiongemma-26B-A4B-it-HERETIC-Uncensored-GGUF:Q4_K_M
Raise the context window
Type it inside the chat. For the API server instead, start it with OLLAMA_CONTEXT_LENGTH=16384 ollama serve.
$/set parameter num_ctx 16384
A larger context needs more memory than the weights alone.
Use the API
Ollama serves its API on http://localhost:11434.

Works on macOS, Windows and Linux.

Works on macOS, Windows and Linux.

Install llama.cpp (macOS)
MacPorts, Nix and conda-forge also work.
$brew install llama.cpp
Install options ↗
Install llama.cpp (Windows)
$winget install llama.cpp
Install llama.cpp (Linux)
Download the prebuilt llama-*-bin-ubuntu-* archive for your GPU backend (CPU, Vulkan) from the releases page, unpack it and run the binaries from that folder. Homebrew (brew install llama.cpp) and conda-forge also work on Linux.
llama.cpp releases ↗
Start the server
Downloads the model on first run. The chat UI and an OpenAI-compatible API are at http://localhost:8080.
$llama-server -hf yabaimimi/edwixx__diffusiongemma-26B-A4B-it-HERETIC-Uncensored-GGUF:Q4_K_M -c 16384 -ngl 99 --port 8080

Works on macOS, Windows and Linux.

Option A: Docker Model Runner
Docker Desktop: Settings > AI > Enable Docker Model Runner. Docker Engine on Linux: install the docker-model-plugin package.
$docker model run hf.co/yabaimimi/edwixx__diffusiongemma-26B-A4B-it-HERETIC-Uncensored-GGUF:Q4_K_M
Model Runner docs ↗
The quant tag after the repo is how Hugging Face cards write it; if your Docker version rejects it, drop the tag and pick the quant by repo.
Option B: llama.cpp server image (NVIDIA)
Needs the NVIDIA Container Toolkit. Models are cached in the llama-cache volume. Open http://localhost:8080.
$docker run --gpus all -p 8080:8080 -v llama-cache:/root/.cache ghcr.io/ggml-org/llama.cpp:server-cuda -hf yabaimimi/edwixx__diffusiongemma-26B-A4B-it-HERETIC-Uncensored-GGUF:Q4_K_M -c 16384 -ngl 99 --host 0.0.0.0 --port 8080
Image tags (cuda13, vulkan, rocm) ↗
Cache path inside the image is not in the docs; remove the -v flag if the volume complains.

Linux only.

Install vLLM
Linux only; on Windows run it inside WSL2. NVIDIA GPU shown; AMD ROCm, Intel and CPU builds have their own install pages.
$uv venv --python 3.12 --seed
$source .venv/bin/activate
$uv pip install vllm --torch-backend=auto
vLLM installation ↗
Serve the model
OpenAI-compatible API at http://localhost:8000/v1.
$vllm serve edwixx/diffusiongemma-26B-A4B-it-HERETIC-Uncensored --max-model-len 16384
Runs the original weights, not the Q4_K_M GGUF; needs ~57 GB of GPU memory.

Works on macOS, Windows and Linux.

Install Jan
jan.ai ↗
Find the model
Open the Hub tab, paste this model ID in Search, and choose Q4_K_M if offered.
$yabaimimi/edwixx__diffusiongemma-26B-A4B-it-HERETIC-Uncensored-GGUF
Use the local API (optional)
Settings > Local API Server > Start Server. It listens on http://127.0.0.1:1337.

Rent GPUs hourly to run Q4_K_M right there. Pricing comes from partner marketplaces and varies.

Available GPU rentals for Q4_K_M, prioritized by value
GPU	Speed	$/1M tokens	Rent
RTX 3090 (24 GB)CheapestCheapest7 GB free	~162 tok/s	$0.34	Deploy on Vast.aifrom $0.20/hr (opens in new tab)	24 GB · 7 GB free · ~162 tok/s · $0.34 / 1M tokens
RTX 4000 Ada (20 GB)3 GB free	~74 tok/s	$0.70	Deploy on Vast.aifrom $0.19/hr (opens in new tab)	20 GB · 3 GB free · ~74 tok/s · $0.70 / 1M tokens
TITAN RTX (24 GB)7 GB free	~126 tok/s	$0.48	Deploy on Vast.aifrom $0.22/hr (opens in new tab)	24 GB · 7 GB free · ~126 tok/s · $0.48 / 1M tokens
RTX A5000 (24 GB)7 GB free	~139 tok/s	$0.46	Deploy on Vast.aifrom $0.23/hr (opens in new tab)	24 GB · 7 GB free · ~139 tok/s · $0.46 / 1M tokens
RTX PRO 4500 (32 GB)15 GB free	~157 tok/s	$0.60	Deploy on RunPodfrom $0.34/hr (opens in new tab)	32 GB · 15 GB free · ~157 tok/s · $0.60 / 1M tokens
RTX 4090 (24 GB)7 GB free	~171 tok/s	$0.55	Deploy on RunPodfrom $0.34/hr (opens in new tab)	24 GB · 7 GB free · ~171 tok/s · $0.55 / 1M tokens
A40 (48 GB)31 GB free	~129 tok/s	$0.75	Deploy on RunPodfrom $0.35/hr (opens in new tab)	48 GB · 31 GB free · ~129 tok/s · $0.75 / 1M tokens
L4 (24 GB)7 GB free	~63 tok/s	$1.93	Deploy on RunPodfrom $0.44/hr (opens in new tab)	24 GB · 7 GB free · ~63 tok/s · $1.93 / 1M tokens

Compare GPUs →

Your own private copy of Diffusiongemma 26B A4B IT HERETIC Uncensored · Q4_K_M · OpenAI-compatible API

Open the llama.cpp listing on RunPod
Open on RunPodServerless (opens in new tab)
New to RunPod? Sign up and add credit first.
2
Fill in the form
Press Deploy and fill in:
Model
yabaimimi/edwixx__diffusiongemma-26B-A4B-it-HERETIC-Uncensored-GGUF
Quantization
Q4_K_M
Context Size
16384
After pasting the model press Esc — don’t pick a suggestion. Press Next.
Create the endpoint
Keep the default 24 GB and 48 GB, then Create Endpoint. Right away: Manage → Edit endpoint → Max workers 1 → Save (changing settings later restarts the worker).
The first start downloads the model, usually 5–15 min. After that, replies after a pause come in seconds.
Create an API key
RunPod → Settings → API Keys → Create. Pick “Restricted” with access to this endpoint only.

Use it

Copy the endpoint URL from its page on RunPod (any of its links works) and paste it with the key. Both stay in this browser.

Endpoint URL

API key

Chats and keys stay only in this browser.

https://api.runpod.ai/v2/<ENDPOINT_ID>/openai/v1

Put your endpoint ID in place of <ENDPOINT_ID>. The API key is the one from step 2.

from openai import OpenAI
client = OpenAI(base_url="https://api.runpod.ai/v2/<ENDPOINT_ID>/openai/v1", api_key="rpa_...")
r = client.chat.completions.create(
    model="any",
    messages=[{"role": "user", "content": "Hello!"}])
print(r.choices[0].message.content)

Works with any OpenAI client.

https://api.runpod.ai/v2/<ENDPOINT_ID>/openai/v1

Put your endpoint ID in place of <ENDPOINT_ID>. The API key is the one from step 2.

Free on Windows, macOS, Linux, iOS, Android and the web.

Open the web app ↗ (opens in a new tab)Download ↗ (opens in a new tab)

1. Settings → Model Provider → Add → Add Custom Provider
2. API Mode OpenAI API Compatible; paste the API key and the base URL as API Host
3. Model → Fetch, pick the model
4. New chat → pick the provider → start chatting

Free and open source for Windows, macOS and Linux.

Download Jan ↗ (opens in a new tab)

1.Settings → next to Model Providers press +
2.API format OpenAI-compatible; Base URL is the base URL above (with /v1); API key is your key; Create
3.If no model appears, click + in Models and type any
4.New chat → pick the model → start chatting