Skip to content

Qwen3.8 Flash-Next on M5 Ultra 512 GB: VRAM, speed and fit

Qwen3.8 Flash-Next at 8-bit runs on the M5 Ultra 512 GB with 256K of context at about 66 tokens a second.

Text generation · Chat · Ollama / LM Studio

Runs comfortably · on M5 Ultra 512 GB · 8-bit (Q8_0) · Ollama / LM Studio · Chat

GPU memory
313 GB
of 512 GB
Context
256K
auto
Speed
66 tok/s
43 when full
First reply
1.0 s
2K prompt

Where the memory goes

Model
178 GB
Context
6.1 GB
Overhead
821 MB
OS
128 GB
Free
199 GB
Memory
313 GB of 512 GB
Context
256K tokens · auto
Speed
66 tok/s · 43 when full
First reply
1.0 s for a 2K prompt
Running cost
$5.06/mo at home · $287/mo on a cloud 4× A40

Validation: Validated

What Qwen3.8 Flash-Next on M5 Ultra 512 GB was checked against
CheckTargetResultStatus
Fits in GPU memory384 GB usable185 GB neededPass
Headroom5% free199 GB free (52%)Pass
Context for chat16K or more256KPass
Writing speed15 tok/s or more66 tok/sPass
First replyunder 3 s, 2K prompt1.0 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
Qwen3.8 Flash-Next
Parameters
180B, 6B active
Precision
8-bit (Q8_0)
Context
256K, auto
Chats at once
1
Use
Chat
Runs in
Ollama / LM Studio
Hardware
M5 Ultra 512 GB
GPU can use
384 GB of 512 GB

Running cost

At home
$5.06/mo
231 W at the wall
Renting instead
$287/mo
4× A40 at $2.36/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

Ollama
ollama run qwen3.8-flash-next:125b-a6b-q4_K_M
llama.cpp
llama-server -hf unsloth/Qwen3.8-Flash-Next-GGUF
vLLM
vllm serve Qwen/Qwen3.8-Flash-Next
MLX (Mac)
mlx_lm.chat --model mlx-community/Qwen3.8-Flash-Next-4bit

Qwen3.8 Flash-Next on similar machines

All machines: Qwen3.8 Flash-Next VRAM requirements. Everything that runs here: what can the M5 Ultra 512 GB run?

Other models that run on M5 Ultra 512 GB

Questions

Can Qwen3.8 Flash-Next run on M5 Ultra 512 GB?

Qwen3.8 Flash-Next at 8-bit runs on the M5 Ultra 512 GB with 256K of context at about 66 tokens a second.

Report VR-1N64QTQ, calculated Nov 22, 2026 from published specs, not benchmarked.