Yes — with 133.7 GB to spare
Qwen3.5 122B-A10B at Q4_K_M fits your CPU only · DDR5 8-channel server entirely on the GPU at 8K context, at an estimated 10 tokens per second. There is room for its full 256K window.
Fully on GPU
8K context
Q4_K_M · 70.3 GB
Apache 2.0
Released 24 Feb 2026
Vision
The 96–128 GB unified-memory model: 122B of knowledge at 10B-active speed.
The VRAM budget
weights 70.3 GB
Weights 70.3 GB
KV cache @ 8K 0.19 GB
Runtime overhead 0.6 GB
Free 133.7 GB of 204.8 GB
Quantisation ladder
| Quant | Weights | Total @ 8K | Max context | Tok/s | Quality | Fit |
|---|---|---|---|---|---|---|
| Q8_0 | 123.7 GB | 124.5 GB | 256K | 5.8 | −0.1% ppl | Long context |
| Q6_K | 95.5 GB | 96.2 GB | 256K | 7.5 | −0.4% ppl | Long context |
| Q5_K_M | 82.5 GB | 83.3 GB | 256K | 8.7 | −0.8% ppl | Long context |
| Q4_K_M | 70.3 GB | 71.1 GB | 256K | 10 | −1.9% ppl | Recommended |
| Q3_K_M | 56.9 GB | 57.7 GB | 256K | 13 | −5.4% ppl | Long context |
| Q2_K | 48.7 GB | 49.5 GB | 256K | 15 | −15% ppl | Long context |
Quality is the published perplexity delta against f16 weights. Max context assumes an f16 KV cache; q8_0 roughly doubles it. Only 12 of its 48 blocks keep a per-token KV cache; the rest are linear-attention, Mamba or convolution blocks with a fixed-size state.
How to run it
$ llama-server \
-hf Qwen/Qwen3.5-122B-A10B:Q4_K_M \
-c 8192 -ngl 99
The engine underneath most of the others. Every knob is exposed. More on llama.cpp.
01Download is 70.3 GB. Keep it on an SSD — a first load off a spinning disk takes minutes.
02With no GPU, thread count matters more than clock. Start at one thread per physical core.
03There is room to go to the model's full 256K context on this card.