Qwen3.8 27B Kearuga — Hardware Requirements & GPU Compatibility
Chat
Qwen3.8 27B Kearuga is a 19.8B-parameter open language model from 0xWhiteMage in the Qwen 3.8 family. It supports a context window of up to 262,144 tokens. At BF16 it needs about 40.27 GB of VRAM — see which GPUs and Macs can run it below.
Qwen3.8 27B Kearuga runs on a 48 GB+ Mac or a workstation GPU
Best local pick → Mac Studio M5 Max (48 GB) · BF16 · 40.7 GB · ~11 tok/s
Macs from 48 GB (BF16 · ~11 tok/s)
Apache 2.0
Not set up for it?
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run Qwen3.8 27B Kearuga
One recommended way. Every other method, OS and quant is in the step-by-step guide.
Recommended
vllm serve 0xWhiteMage/Qwen3.8-27B-Kearuga --max-model-len 16384vLLM repo with its own quantisation
Download: Weights on Hugging Face
Other ways
each opens step-by-stepRun in cloud
Fits on A100 80GB (39 GB headroom) · BF16
- Generation speed
- ~33 tok/s
- generation speed
- Cost per 1M output tokens
- $9.77
- per 1M output tokens
How Much VRAM Does Qwen3.8 27B Kearuga Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 40.3 GB | 41.6 GB | 16-bit · bfloat — Brain floating point 16 — preferred for training | Does not fit: needs 40.7 GB |
Selected BF16
Brain floating point 1640.7GBat 16K context
- Weights
- 39.5 GB
- KV cache
- 0.9 GB
- Overhead
- 0.3 GB
Doesn't fit RTX 4090 (reference) · needs 16.7 GB more
Which Hardware Can Run Qwen3.8 27B Kearuga?
BF16 · 40.3 GBQwen3.8 27B Kearuga (BF16) requires 40.3 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 53+ GB is recommended. Using the full 262K context window can add up to 13.9 GB, bringing total usage to 54.1 GB. No consumer GPU has enough memory. 67 devices with unified memory can run Qwen3.8 27B Kearuga, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), Mac Studio M5 Max (48 GB).
48 GB
48 GB tier →fits · 7.3 GB free · up to 128K ctx
2 Mac classes · 1 device family
M5 Max48 GB~11Simulate output on Mac Studio M5 Max (48 GB)M5 Pro48 GB~5Simulate output on Mac Mini M5 Pro (48 GB)Snapdragon X2 Elite Extreme (X2E-96-100)~4Simulate output on Snapdragon X2 Elite Extreme Copilot+ PC
By VRAM tier · typical GPUs and Macs at BF16
Alternatives to Qwen3.8 27B Kearuga
Qwen 3.8 family · 12 models →Similar models
5| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| InternVL3 5 GPT OSS 20B A4B Preview HFOpenGVLab | — | 13.6 GB | ~180 | same size |
| Internlm 20BInternLM | — | 45.3 GB | —needs > 24 GB | same size |
| Tri 21B Thinktrillionlabs | — | 45.1 GB | —needs > 24 GB | same size |
| ERNIE 4.5 21B A3B PTBaidu | — | 13.8 GB | ~254 | same size |
| GPT Neox 20BEleutherAI | — | 13.7 GB | ~48 | same size |
Specifications
- Publisher
- 0xWhiteMage
- Family
- Qwen 3.8
- Parameters
- 19.8B
- Architecture
- Qwen3_5ForConditionalGeneration
- Context Length
- 262,144 tokens
- Vocabulary Size
- 248,320
- Release Date
- 2026-09-04
- License
- Apache 2.0
Frequently Asked Questions
How much VRAM does Qwen3.8 27B Kearuga need?
Qwen3.8 27B Kearuga requires 40.3 GB of VRAM at BF16. Full 262K context adds up to 13.9 GB (54.1 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 19.8B × 16 bits ÷ 8 = 39.5 GB
KV Cache + Overhead ≈ 0.8 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 ≈ 14.6 GB (at full 262K context)
VRAM usage by quantization
Can NVIDIA GeForce RTX 5090 run Qwen3.8 27B Kearuga?
No — Qwen3.8 27B Kearuga requires at least 40.3 GB at BF16, which exceeds the NVIDIA GeForce RTX 5090's 32 GB of VRAM.
Can I run Qwen3.8 27B Kearuga on a Mac?
Yes — Mac Studio M5 Max (48 GB) and 55 other Macs can run Qwen3.8 27B Kearuga. Apple Silicon uses unified memory, so the model shares RAM with the system. At BF16 you need at least 40.3 GB of usable unified memory (RAM minus macOS overhead).
Can I run Qwen3.8 27B Kearuga locally?
Yes — Qwen3.8 27B Kearuga can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is Qwen3.8 27B Kearuga?
At BF16, Qwen3.8 27B Kearuga can reach ~119 tok/s on AMD Instinct MI350X. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: NVIDIA B200 → 8000 ÷ 40.3 × 0.65 = ~129 tok/s
Estimated speed at BF16 (40.3 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of Qwen3.8 27B Kearuga?
At BF16, the download is about 39.52 GB.
Which GPUs can run Qwen3.8 27B Kearuga?
No single consumer GPU has enough VRAM to run Qwen3.8 27B Kearuga at BF16 (40.3 GB). Multi-GPU or professional hardware is required.
Which devices can run Qwen3.8 27B Kearuga?
67 devices with unified memory can run Qwen3.8 27B Kearuga at BF16 (40.3 GB), including ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (Ryzen AI Max+ 395, 128 GB), GMKtec EVO-X2 (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.