back to local model checker

Qwen 3.5 122B-A10B

MoEThinkingVision

Alibaba · 122B (10B active) · Mixture of Experts

Large multimodal MoE

Apache 2.0
Downloads

253.4K

Likes

411

Released

2026-02

Context

256K tokens

Use cases

chatvisionreasoning
Total experts256
Active experts8
Active params10.0B

Your hardware

16 GB Reference Desktop

Quant
Q4_K_M
Memory
63 GB
Speed
--
Fit
Too heavy

Quantization options

What is quantization?
QuantBitsVRAMDiskQualityStatus
Q4_K_M463 GB37.3 GBgoodToo heavy · ~1 t/s
Q6_K694.2 GB37.1 GBexcellentToo heavy · ~1 t/s
Q8_08125.5 GB37.2 GBexcellentToo heavy · ~1 t/s

About this model

Qwen 3.5 122B-A10B (Qwen3.5 122b) is a 122B sparse mixture-of-experts (MoE) model activating 10B active per token from Alibaba in the Qwen family, designed primarily for chat, vision, reasoning workflows. Native reasoning/thinking tokens expand output length per query, so reserve extra KV cache headroom. Supports up to 256K context tokens; start at 4K–8K with Q8_0 KV cache on tight GPUs.

Hardware planning baselines start at 68.2 GB minimum memory and 113.6 GB recommended memory. At Q4_K_M, weights plus baseline runtime overhead use roughly 63 GB of RAM/VRAM and 37.3 GB of disk space. At 122B, Qwen 3.5 122B-A10B belongs firmly in the workstation and multi-GPU tier. Running it comfortably requires 113.6 GB of total memory across dual RTX GPUs, high-memory Apple Silicon (Mac Studio Ultra/Max), or MoE hybrid CPU-RAM offloading—mobile devices and standard 16 GB laptops are too constrained for practical speeds.

Deploy Qwen 3.5 122B-A10B locally via Ollama (`qwen3.5:122b-a10b`), Ollama with Flash Attention enabled. Under Apache License 2.0, review official weight cards before commercial redistribution.

Highlights

  • Local Fit Baseline68.2 GB minimum and 113.6 GB recommended RAM/VRAM.
  • Primary QuantQ4_K_M targets 63 GB memory and 37.3 GB disk.
  • Context & Capabilities256K max context · Sparse MoE · Reasoning.

System requirements by device

Memory planning targets 68.2 GB minimum RAM/VRAM and 113.6 GB recommended memory. The baseline quant is Q4_K_M, estimated at 63 GB memory and 37.3 GB disk. Actual runtime usage shifts with context length, KV cache precision, and OS background overhead.

DeviceNotes
Mac (Apple silicon)Requires a 48 GB–192 GB Mac Studio or MacBook Pro Max/Ultra with enough unified wired budget for 63 GB.
Windows / Linux GPUUse dual/multi-GPU CUDA splitting (--tensor-split) or MoE CPU expert offloading (-ot) paired with 32–128 GB DDR5 system RAM.
Phones & Compact Edge BoardsNot practical on phones (68.2 GB min RAM exceeds Android LMK and iOS Jetsam limits). Pick a 2B–4B model on mobile.

Start with a short session and leave memory for the operating system. A file fitting on disk does not guarantee that it will load in memory.

Estimated memory by quantization

These are rough estimates, not tested minimum requirements. Check the actual file and runtime documentation; context, offloading and operating-system overhead can change the memory needed.

Quant TierEstimated memoryDisk SizeQuality
Q4_K_M63.0 GB37.3 GBgood
Q6_K94.2 GB37.1 GBexcellent
Q8_0125.5 GB37.2 GBexcellent

How to run Qwen 3.5 122B-A10B on your device

Start with a short prompt and the smallest quant that still gives the quality you need. A clean first run is more useful than loading a larger file with no memory headroom.

Windows

Run Qwen 3.5 122B-A10B on Multi-GPU or High-RAM Desktop

  1. Plan for 68.2 GB minimum combined memory and 113.6 GB recommended memory for Q4_K_M.
  2. Use llama.cpp `-ot ".ffn_.*_exps.=CPU"` or `--tensor-split` across multiple GPUs so attention and KV cache remain in fast VRAM.
  3. Enable Flash Attention (`-fa on`) and Q8_0 KV cache (`-ctk q8_0 -ctv q8_0`) to prevent context buffers from exceeding GPU headroom.
  4. Avoid attempting to run workstation-scale weights on smartphones or 8 GB laptops; use a 2B–8B checkpoint on portable devices.
Device how-to guide
Mac

Run Qwen 3.5 122B-A10B on High-Memory Apple Silicon Mac

  1. Verify you are on a 48 GB–192 GB Mac Studio or MacBook Pro Max/Ultra capable of holding 63 GB in wired unified memory.
  2. Launch via `llama-server` or LM Studio with MLX/Metal enabled and set initial context conservatively to 4K–8K.
  3. Monitor Activity Monitor Memory Pressure to confirm zero SSD swap thrashing during token generation.
  4. Step down to a lighter quantization (e.g., Q3_K_M or IQ4_XS) if total memory utilization exceeds 75%.
Device how-to guide

Qwen 3.5 122B-A10B vs alternatives

ModelParamsContextMin RAMLicense
Qwen 3.5 122B-A10B122B (10B active)256K68.2 GBApache 2.0
Qwen 3.8 Flash 128B125B (6B active)128K32 GBApache 2.0
Qwen 3 Next 80B-A3B80B (3B active)256K44.7 GBApache 2.0
Qwen 3 Coder Next 80B-A3B80B (3B active)256K44.7 GBApache 2.0
Qwen 2.5 72B72B128K40.2 GBQwen
Qwen 3.6 35B-A3B36B (3B active)256K20.1 GBApache 2.0

Use this table to choose the smallest nearby model that still matches the job. A related model with lower RAM can be the better local choice when the current page fits only at the edge of available memory.

Frequently asked questions

What is Qwen 3.5 122B-A10B?
Qwen 3.5 122B-A10B is a 122B sparse mixture-of-experts model (10B active active per token) from Alibaba, supporting up to 256K tokens with 68.2 GB minimum RAM and 113.6 GB recommended RAM.
What is the best way to run Qwen3.5 122b locally?
Choose the smallest quant that matches the task, test one short prompt, then compare RAM, context, and nearby models before making Qwen 3.5 122B-A10B your default local model.
How much VRAM or RAM does Qwen 3.5 122B-A10B require?
At Q4_K_M, Qwen 3.5 122B-A10B occupies roughly 63 GB of active memory and 37.3 GB on disk. Add 2–6 GB of spare headroom for KV cache buffers, runtime scratch space, and OS apps (113.6 GB recommended).
What is the best way to run Qwen 3.5 122B-A10B locally?
Use `ollama run qwen3.5:122b-a10b` for a quick CLI/API setup, or load Q4_K_M in LM Studio or llama.cpp with Flash Attention enabled when you want explicit KV cache and layer control.
Can Qwen 3.5 122B-A10B run on a consumer laptop or smartphone?
No smartphone or 8–16 GB laptop can hold Qwen 3.5 122B-A10B cleanly in local memory. It targets multi-GPU NVIDIA workstations, high-RAM Mac Studios, or MoE CPU-RAM hybrid rigs with at least 68.2 GB total memory.
How does Qwen 3.5 122B-A10B compare to Qwen 3.8 Flash 128B?
Qwen 3.8 Flash 128B (125B, 32 GB min RAM) sits closest in size. Compare active parameters, context ceiling (256K vs 128K), and license terms (Apache 2.0) before switching.