KRAFTONRaon1MiniForCausalLM

Raon 1 Mini Preview — Hardware Requirements & GPU Compatibility

ChatFunctions

Raon 1 Mini Preview is a 28.8B-parameter open language model from KRAFTON. It supports a context window of up to 262,144 tokens. At Q4_K_M it needs about 19.00 GB of VRAM — see which GPUs and Macs can run it below.

28.8B params · MoE4.3B active262K contextApache 2.0Oct 20263.3K downloads228 likes37.2K quant downloads

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.

Decent on a 24 GB GPU

Runs on 24 GB (Q4_K_M · ~175 tok/s · 16K)Macs from 24 GB (Q4_K_M · ~55 tok/s)

Apache 2.0

Reference GPU · RTX 4090 24 GB

Estimates, ±20% · how we estimate

Run Raon 1 Mini Preview

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.

Ways to runOther ways

each opens step-by-step

Run in cloud

Runs on RTX A6000 (48 GB) (25 GB headroom) · Q4_K_M

Generation speed: ~141 tok/s; generation speed
Cost per 1M output tokens: $0.65; per 1M output tokens

Deployon RunPodBest valuefrom $0.33/hr

$5+ bonus for new users

Compare GPUs →

Run Raon 1 Mini Preview

22.7 GB VRAM at 16K context · fits the RTX 4090 (1.3 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/KRAFTON/Raon-1-mini-preview-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.

Install LM Studio
Desktop app for macOS (Apple silicon), Windows and Linux.
lmstudio.ai ↗
Download the model
In the app press Ctrl/Cmd + Shift + M and paste https://huggingface.co/KRAFTON/Raon-1-mini-preview-gguf, then pick Q4_K_M. Or use the CLI (run LM Studio once first):
$lms get https://huggingface.co/KRAFTON/Raon-1-mini-preview-gguf@Q4_K_M
Open on Hugging Face ↗
Load and chat
Open the model loader, select the model and set Context Length to 16384. Leave GPU offload at maximum if it fits.
Start the local server (optional)
$lms server start

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 KRAFTON/Raon-1-mini-preview-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/KRAFTON/Raon-1-mini-preview-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 KRAFTON/Raon-1-mini-preview-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 KRAFTON/Raon-1-mini-preview --max-model-len 16384
Runs the original weights, not the Q4_K_M GGUF; needs ~63 GB of GPU memory.

macOS only.

MLX supports only macOS, so follow these steps for that setup.

Install mlx-lm
$pip install mlx-lm
Chat in the terminal
$mlx_lm.chat --model KRAFTON/Raon-1-mini-preview-mlx
Downloads the model on first run. The MLX repo has its own quantisation; it may not match the quant selected above.
Or start a local server
Listens on http://localhost:8080.
$mlx_lm.server --model KRAFTON/Raon-1-mini-preview-mlx

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.
$KRAFTON/Raon-1-mini-preview-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 A6000 (48 GB)CheapestCheapest25 GB free	~141 tok/s	$0.65	Deploy on RunPodfrom $0.33/hr (opens in new tab)	48 GB · 25 GB free · ~141 tok/s · $0.65 / 1M tokens
RTX PRO 4500 (32 GB)9 GB free	~160 tok/s	$0.59	Deploy on RunPodfrom $0.34/hr (opens in new tab)	32 GB · 9 GB free · ~160 tok/s · $0.59 / 1M tokens
A40 (48 GB)25 GB free	~130 tok/s	$0.75	Deploy on RunPodfrom $0.35/hr (opens in new tab)	48 GB · 25 GB free · ~130 tok/s · $0.75 / 1M tokens
RTX 5000 Ada (32 GB)9 GB free	~111 tok/s	$1.11	Deploy on Vast.aifrom $0.44/hr (opens in new tab)	32 GB · 9 GB free · ~111 tok/s · $1.11 / 1M tokens
RTX A6000 (48 GB)25 GB free	~141 tok/s	$0.90	Deploy on Vast.aifrom $0.46/hr (opens in new tab)	48 GB · 25 GB free · ~141 tok/s · $0.90 / 1M tokens
L40 (48 GB)25 GB free	~155 tok/s	$0.91	Deploy on Vast.aifrom $0.51/hr (opens in new tab)	48 GB · 25 GB free · ~155 tok/s · $0.91 / 1M tokens
A40 (48 GB)25 GB free	~130 tok/s	$1.14	Deploy on Vast.aifrom $0.54/hr (opens in new tab)	48 GB · 25 GB free · ~130 tok/s · $1.14 / 1M tokens
RTX 5090 (32 GB)9 GB free	~262 tok/s	$0.73	Deploy on RunPodfrom $0.69/hr (opens in new tab)	32 GB · 9 GB free · ~262 tok/s · $0.73 / 1M tokens

Compare GPUs →

How Much VRAM Does Raon 1 Mini Preview Need?

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

VRAM	+32K	+128K	File	Quality	On RTX 4090 (reference)
13.5 GB	21.3 GB	46.5 GB	~12.2 GB	2-bit · K-quant — 2-bit quantization with K-quant improvements	Fits: 32K context · ~224 tok/s
15.4 GB	23.3 GB	48.5 GB	~14 GB	3-bit medium — 3-bit medium quantization	Fits: 32K context · ~204 tok/s
17.0 GB	24.9 GB	50.1 GB	16.1 GB	4-bit · i-quant compact — Importance-weighted 4-bit, compact	Fits: 16K context · ~190 tok/s
19 GB	26.9 GB	52.1 GB	19.7 GB	4-bit medium — 4-bit medium quantization — most popular sweet spot	Fits: 16K context · ~175 tok/s
22.6 GB	30.4 GB	55.6 GB	22.2 GB	5-bit medium — 5-bit medium quantization — good quality/size tradeoff	Fits: 4K context · ~153 tok/s
26.1 GB	34 GB	59.2 GB	25.8 GB	6-bit, very good — 6-bit quantization, very good quality	Does not fit: needs 29.8 GB
31.7 GB	39.5 GB	64.7 GB	~28.8 GB	8-bit, near-lossless — 8-bit quantization, near-lossless	Does not fit: needs 35.3 GB
63.3 GB	71.2 GB	96.4 GB	~57.6 GB	16-bit · bfloat — Brain floating point 16 — preferred for training	Does not fit: needs 67 GB

est.calculated, no published file yet

Selected Q4_K_M

4-bit medium quantization

22.7GBat 16K context

Weights: 18.2 GB
KV cache: 4.2 GB
Overhead: 0.3 GB

Fits RTX 4090 (reference) · 1.3 GB free · ~175 tok/s

≥50 tok/s needs ≥235 GB/s · RTX 5090 ~262, RTX 3090 Ti ~175

Which Hardware Can Run Raon 1 Mini Preview?

Q4_K_M · 19 GB

Raon 1 Mini Preview (Q4_K_M) requires 19 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 25+ GB is recommended. 8 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti. 122 devices with unified memory can run Raon 1 Mini Preview, 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

8–16 GB · RTX 4060, RTX 3060 12GB, RTX 5060 Ti 16GBDoesn't fit — smallest quant needs 17.1 GB8 GB tier →12 GB tier →16 GB tier →

24 GB · RTX 4090referenceQ4_K_M · 22.7 GB · ~175 tok/s · 16K22.7~17516K24 GB tier →

Mac 32 GB · Mac Mini M4Q4_K_M · 22.7 GB · ~26 tok/s · 32K22.7~2632KMacs →

Mac 64 GB · Mac Studio M4 MaxQ4_K_M · 22.7 GB · ~80 tok/s · 128K22.7~80128KMacs →

Alternatives to Raon 1 Mini Preview

Similar models

6tok/s on RTX 4090

Model	Score	VRAM	tok/s	Why it is similar
Qwen3.8 27BAlibaba	121.9	17.9 GB	~38	same size
Qwen3.6 27BAlibaba	111.2	17.9 GB	~38	same size
Hy Memory 30B A3BTencent	—	19.1 GB	~184	same size
Granite Switch 4.1 30BIBM	—	22.7 GB	~35	same size
Muse Glimmer 30BMeta	—	19.6 GB	~36	same size
Hy MT2 30B A3BTencent	—	19.1 GB	~170	same size

Fine-tunes & variants(1)community builds of Raon 1 Mini Preview

Raon 1 Mini Preview Dflash2

Specifications

Publisher: KRAFTON
Parameters: 28.8B
Active parameters: 4.3B per token
Architecture: Raon1MiniForCausalLM
Context Length: 262,144 tokens
Vocabulary Size: 259,840
Release Date: 2026-10-21
License: Apache 2.0

Resources

Hugging Face model card (opens in a new tab)

Frequently Asked Questions

How much VRAM does Raon 1 Mini Preview need?

Raon 1 Mini Preview requires 19 GB of VRAM at Q4_K_M, or 63.3 GB at BF16.

VRAM = Weights + KV Cache + Overhead

Weights = 28.8B × 4.8 bits ÷ 8 = 17.3 GB

KV Cache + Overhead ≈ 1.7 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.

VRAM usage by quantization

Q4_K_M

19.0 GB

Learn more about VRAM estimation →

Can NVIDIA GeForce RTX 4090 run Raon 1 Mini Preview?

Yes, at Q5_K_M (22.6 GB) or lower. Higher quantizations like Q6_K (26.1 GB) exceed the NVIDIA GeForce RTX 4090's 24 GB.

What's the best quantization for Raon 1 Mini Preview?

For Raon 1 Mini Preview, Q4_K_M (19 GB) offers the best balance of quality and VRAM usage. Q5_K_M (22.6 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 13.5 GB.

VRAM requirement by quantization

Q2_K

13.5 GB~75%

IQ4_XS

17.0 GB~87%

Q4_K_M ★

19.0 GB~89%

Q5_K_M

22.6 GB~92%

Q6_K

26.1 GB~95%

BF16

63.3 GB~100%

★ Recommended — best balance of quality and VRAM usage.

Learn more about quantization →

Can I run Raon 1 Mini Preview on a Mac?

Yes — Mac Mini M5 Pro (24 GB) and 106 other Macs can run Raon 1 Mini Preview. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 19 GB of usable unified memory (RAM minus macOS overhead).

Can I run Raon 1 Mini Preview locally?

Yes — Raon 1 Mini Preview can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.

How fast is Raon 1 Mini Preview?

At Q4_K_M, Raon 1 Mini Preview can reach ~179 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~175 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) = 0.548 ms, plus 25 layers × 0.055 ms = 1.375 ms, so 1000 ÷ 1.923 ms = ~520 tok/s

Estimated speed at Q4_K_M (19 GB)

NVIDIA B200

~520 tok/s

NVIDIA GeForce RTX 4090

~175 tok/s

NVIDIA B300

~520 tok/s

NVIDIA GH200 Grace Hopper Superchip

~441 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 Raon 1 Mini Preview?

At Q4_K_M, the download is about 17.27 GB. The full-precision BF16 version is 57.57 GB. The smallest option (Q2_K) is 12.23 GB.

Which GPUs can run Raon 1 Mini Preview?

8 consumer GPUs can run Raon 1 Mini Preview 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 Raon 1 Mini Preview?

122 devices with unified memory can run Raon 1 Mini Preview at Q4_K_M (19 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.