# AMD Ryzen AI Max+ 395 (128GB) for local LLMs

> The cheapest way to run 100B-class MoE models in one quiet box, though prices have climbed past $3,300.

The Ryzen AI Max+ 395 has 128GB of unified memory (about 96GB usable by the GPU by default) and 256 GB/s of memory bandwidth, which decides how big a model fits and how fast it generates text. It comfortably runs models up to about 122B parameters: the largest in our catalog is Qwen3.5 122B-A10B at Q4_K_M (77GB with 8K context, ~18 tokens/sec).

## Key specs

- Memory: 128GB LPDDR5X-8000 (unified; ~96GB usable by the GPU by default)
- Memory bandwidth: 256 GB/s
- Architecture: RDNA 3.5 (Strix Halo) (2025)
- Power: 120W
- Typical price: ~$3,300 (checked 2026-10)
- Software: ROCm, Vulkan

## Best models for this GPU

- **Best overall (also coding, reasoning, vision, largest that fits):** Qwen3.5 122B-A10B at Q4_K_M, ~77.0GB with 8K context, ~18 tokens/sec
- **Runner-up:** Qwen3.6 35B-A3B at Q8_0, ~41.0GB with 8K context, ~30 tokens/sec
- **Fastest good model:** gpt-oss-20b at MXFP4, ~15.0GB with 8K context, ~41 tokens/sec

## Compatibility (best quantization that fits, 8K context)

| Model | Params | Quant | Memory | Speed | Verdict |
| --- | --- | --- | --- | --- | --- |
| Gemma 4 E4B | 8B | Q8_0 | 10.5GB | ~20 tok/s | Runs well |
| Qwen3.5 9B | 9B | Q8_0 | 12.0GB | ~17 tok/s | Runs well |
| Phi-4 | 14.7B | Q8_0 | 18.3GB | ~12 tok/s | Runs well |
| gpt-oss-20b | 20.9B (3.6B active) | MXFP4 | 15.0GB | ~41 tok/s | Runs well |
| Gemma 4 26B A4B | 26B (3.8B active) | Q8_0 | 29.5GB | ~26 tok/s | Runs well |
| Qwen3.8 27B | 27B | Q8_0 | 31.6GB | ~6.4 tok/s | Runs well |
| Gemma 4 31B | 31B | Q8_0 | 35.9GB | ~5.8 tok/s | Runs well |
| Qwen3.6 35B-A3B | 35B (3B active) | Q8_0 | 41.0GB | ~30 tok/s | Runs well |
| GLM-4.5-Air | 106B (12B active) | Q5_K_M | 79.9GB | ~13 tok/s | Runs well |
| Llama 4 Scout | 109B (17B active) | Q5_K_M | 81.3GB | ~9.3 tok/s | Runs well |
| gpt-oss-120b | 117B (5.1B active) | MXFP4 | 68.3GB | ~35 tok/s | Runs well |
| Qwen3.5 122B-A10B | 122B (10B active) | Q4_K_M | 77.0GB | ~18 tok/s | Runs well |
| DeepSeek V4 Flash | 284B (13B active) | - | - | - | Won't fit |
| GLM-5.3-Flash | 320B (18B active) | - | - | - | Won't fit |
| GLM-5.3 | 744B (40B active) | - | - | - | Won't fit |

## Pros
- Up to 96GB of GPU memory on Windows and about 120GB on Linux
- Small, quiet systems with configurable 45 to 120W chip power
- Strong for MoE models such as gpt-oss-120b
- A full x86 PC that also works as a desktop

## Cons
- 256 GB/s makes dense 70B models slow
- Slower prompt processing than DGX Spark or NVIDIA GPUs
- Prices have risen sharply

Full page: https://llamabuilds.ai/gpus/ryzen-ai-max-plus-395 · Best models: https://llamabuilds.ai/gpus/ryzen-ai-max-plus-395/best-models

Sources: [ComputingForGeeks Ryzen AI Max+ 395 mini PC price comparison](https://computingforgeeks.com/ryzen-ai-max-395-mini-pc-comparison/), [OpenClawDC Strix Halo mini PC buying guide](https://openclawdc.com/blog/which-strix-halo-mini-pc-to-buy/), [Compute Market Strix Halo 128GB prices and benchmarks](https://www.compute-market.com/blog/strix-halo-mini-pc-local-ai-2026)

---
Source: https://llamabuilds.ai/gpus/ryzen-ai-max-plus-395 · Updated October 14, 2026 · Llama Builds. Speeds marked ~ are estimates from our bandwidth model (https://llamabuilds.ai/methodology); measured results are labeled as such.