ai.canshu.com
中文
VRAM Calculator Advisor & TCO API Pricing Compare Models Leaderboard GPUs

Search models, GPUs, and tools

↑↓ Navigate · Enter Open · ★ Watchlist prioritized ai.canshu.com
Q4_K_M: Smooth Fit 19.33 GB / 99.8 GB Usable

Qwen3.5-27B on AMD Ryzen AI Max+ 395 Mini PC (128GB Unified): VRAM Requirement & Speed Estimate

Yes. At Q4_K_M 4-bit quantization with a 16K context window, Qwen3.5-27B requires 19.33 GB of VRAM (16.37 GB weights + 1.34 GB KV cache + 1.63 GB overhead), fitting within AI Max+ 395 128G's 99.8 GB usable memory at ~9.9 tokens/s.

Concurrency (Batch):
Dense 27B·64L (d=5120)·24Q/4KV·Ctx 256K·Text+Image+Video
Runnable Models →
128 GB Unified·256 GB/s·Safe Usable 99.8 GB
Smooth Fit19.33 / 99.8 GB~9.9 tok/s
3. Weight Quantization
4. Per-Request Context16K (16,384 tok)
Verified: 2026-11-20·Roofline bandwidth/compute estimate (Batch=1)
Smooth Fit
19.33/ 99.8 GB usable

Physical 128 GB · Load 19.4%

Theoretical Decode Speed
~9.9tokens/s

Actual speed varies ~±25% by runtime

Max Context at Batch=1
256K/ 256K

Enable Q8/Q4 KV quantization to scale batch concurrency

Weights: 16.37 GBKV Cache (1×16K): 1.34 GBOverhead: 1.63 GB
Reserves ~22% unified memory for OS
🚶 Readable Pace~50.5s for a 500-token response
Reference Launch CLI:
OLLAMA_CONTEXT_LENGTH=16384 ollama run qwen3.5:27b

Insufficient Local VRAM? Ways to Run Qwen3.5-27B

Rent an hourly high-VRAM cloud GPU pod, or check the 12-month Buy vs Cloud vs API break-even curve.

📊 All Quantization Levels for Qwen3.5-27B on AI Max+ 395 128G

Quantization Total VRAM (16K) Verdict Est. Speed Max Context
FP16 / BF16 57.72 GB Smooth Fit ~3.1 tok/s 256K
FP8 / Q8_0 31.9 GB Smooth Fit ~5.8 tok/s 256K
GGUF Q6_K 25.22 GB Smooth Fit ~7.4 tok/s 256K
GGUF Q5_K_M 22.23 GB Smooth Fit ~8.5 tok/s 256K
Q4_K_M / AWQ 19.33 GB Smooth Fit ~9.9 tok/s 256K
GGUF IQ4_XS 17.27 GB Smooth Fit ~11.3 tok/s 256K
GGUF Q3_K_M 16.1 GB Smooth Fit ~12.2 tok/s 256K
GGUF Q2_K 13.48 GB Smooth Fit ~15 tok/s 256K

💡 Recommended Alternative GPUs & Macs for Qwen3.5-27B

Frequently Asked Questions

Can AMD Ryzen AI Max+ 395 Mini PC (128GB Unified) run Qwen3.5-27B locally?

Yes. At Q4_K_M 4-bit quantization with a 16K context window, Qwen3.5-27B requires 19.33 GB of VRAM (16.37 GB weights + 1.34 GB KV cache + 1.63 GB overhead), fitting within AI Max+ 395 128G's 99.8 GB usable memory at ~9.9 tokens/s.

What is the best quantization for Qwen3.5-27B on AI Max+ 395 128G?

The highest-quality quantization that fits cleanly on AI Max+ 395 128G is FP16 / BF16 (Lossless Full Precision), using 57.72 GB VRAM at ~3.1 tokens/s with up to 256K context.

How much VRAM does Qwen3.5-27B need in full FP16 precision?

In unquantized FP16/BF16 precision (16K context), Qwen3.5-27B requires 57.72 GB VRAM (54 GB for weights alone).