Strix HaloAvailable
GMKtec EVO-X2
Ryzen AI Max+ 395, 64GB / 1TB
$2,200at GMKtec store
256 GB/s
Check priceMemory guide · 64GB unified memory
What can you run on 64GB unified memory?
9 of the 16 open models we track fit in 64GB unified memory without spilling into system RAM, the biggest being Llama 3.3 70B. Speeds range from about 5 to 48 tokens/sec depending on the model and machine. Prices & stock checked 2026-10-23. Prices change daily and some retailers block our checker, so always confirm on the store page.
Models that fit in 64GB unified memory
| Model | Needs at 4-bit | GMKtec EVO-X2256 GB/s |
|---|---|---|
| Qwen3.5 9Bdense · 9B active · chat, small agents | ~7 GB | Fits~21 tok/sQ8_0 |
| Gemma 4 12Bdense · 12B active · chat, vision | ~9 GB | Fits~16 tok/sQ8_0 |
| GPT-OSS 20BMoE · 3.6B active · reasoning, tools | ~13 GB | Fits~42 tok/sQ8_0 |
| Gemma 4 26B-A4BMoE · 3.8B active · fast general chat | ~18 GB | Fits~40 tok/sQ8_0 |
| Qwen3.8 27Bdense · 27B active · coding, vision, general | ~19 GB | Fits~12 tok/sQ4_K_M |
| GLM-4.7-Flash 30B-A3BMoE · 3B active · coding agents | ~21 GB | Fits~48 tok/sQ8_0 |
| Gemma 4 31Bdense · 31B active · general, vision | ~21 GB | Fits~11 tok/sQ4_K_M |
| Qwen3.6 35B-A3BMoE · 3B active · fast coding & chat | ~24 GB | Fits~48 tok/sQ8_0 |
| Llama 3.3 70Bdense · 70B active · general | ~46 GB | Tight fit~5 tok/sQ4_K_M |
Each column shows the quant we'd use (8-bit when it fits comfortably and still runs 10+ tok/s, otherwise 4-bit) and the estimated decode speed for one user at ~8K context. How we estimate
Just out of reach: Qwen3-Next 80B-A3B (~52 GB at 4-bit), Llama 4 Scout 109B-A17B (~71 GB at 4-bit), GPT-OSS 120B (~67 GB at 4-bit).
1 tracked rig with 64GB unified memory
FAQ
What's the biggest model I can run on 64GB unified memory?
Of the models we track, Llama 3.3 70B is the largest that fits fully in 64GB unified memory (Q4_K_M, about 46 GB including runtime and ~8K context).
What's the cheapest rig with 64GB unified memory?
Of the machines we track, the GMKtec EVO-X2 at $2,200 (GMKtec store, checked 2026-10-23).
Can I run Llama 3.3 70B on 64GB unified memory?
Yes, at 4-bit it needs about 46 GB including runtime and ~8K context.
How are fit and speed estimated?
Memory needed = weights + 1.5 GB + 5% of weights. Decode speed ≈ 1 ÷ (active weights in GB ÷ (efficiency × memory bandwidth) + a small fixed per-token overhead), with efficiency and overhead fitted per platform to 43 public llama.cpp/MLX benchmarks (median error about 5%). They're estimates, not benchmarks.