back to local model checker

Qwen 3.8 2.4T-A95B

MoEToolsThinkingPopular

Alibaba · 2.4T (95B active) · Mixture of Experts

Frontier Qwen 3.8 MoE — 95B active, 1M context

Qwen
Downloads

0

Likes

0

Released

2026-08

Context

1M tokens

Use cases

chatreasoningcode

Your hardware

16 GB Reference Desktop

Quant
Q4_K_M
Memory
1229.8 GB
Speed
--
Fit
Too heavy

Quantization options

What is quantization?
QuantBitsVRAMDiskQualityStatus
Q2_K2768.8 GB733.4 GBlowToo heavy · ~1 t/s
Q3_K_M31076.2 GB1026.8 GBmoderateToo heavy · ~1 t/s
Q4_K_M41229.8 GB1173.5 GBgoodToo heavy · ~1 t/s
Q5_K_M51537.2 GB1466.8 GBgoodToo heavy · ~1 t/s
Q6_K61844.5 GB1760.2 GBexcellentToo heavy · ~1 t/s
Q8_082459.2 GB2346.9 GBexcellentToo heavy · ~1 t/s
F16164917.9 GB4693.9 GBlosslessToo heavy · ~1 t/s

About this model

Qwen 3.8 2.4T-A95B (Qwen 3.8 2.4t) is a 2.4T sparse mixture-of-experts (MoE) model activating 95B active per token from Alibaba in the Qwen family, designed primarily for chat, reasoning, code workflows. Native reasoning/thinking tokens expand output length per query, so reserve extra KV cache headroom. Supports structured tool calling; use Q4_K_M or higher when strict JSON schema adherence matters. Supports up to 1M context tokens; start at 4K–8K with Q8_0 KV cache on tight GPUs.

Hardware planning baselines start at 1341.1 GB minimum memory and 2235.2 GB recommended memory. At Q4_K_M, weights plus baseline runtime overhead use roughly 1229.8 GB of RAM/VRAM and 1173.5 GB of disk space. At 2.4T, Qwen 3.8 2.4T-A95B belongs firmly in the workstation and multi-GPU tier. Running it comfortably requires 2235.2 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 3.8 2.4T-A95B locally via llama.cpp with Flash Attention enabled. Under Tongyi Qianwen License Agreement, review official weight cards before commercial redistribution.

Highlights

  • Local Fit Baseline1341.1 GB minimum and 2235.2 GB recommended RAM/VRAM.
  • Primary QuantQ4_K_M targets 1229.8 GB memory and 1173.5 GB disk.
  • Context & Capabilities1M max context · Sparse MoE · Tool calling · Reasoning.

System requirements by device

Memory planning targets 1341.1 GB minimum RAM/VRAM and 2235.2 GB recommended memory. The baseline quant is Q4_K_M, estimated at 1229.8 GB memory and 1173.5 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 1229.8 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 (1341.1 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
Q2_K768.8 GB733.4 GBlow
Q3_K_M1076.2 GB1026.8 GBmoderate
Q4_K_M1229.8 GB1173.5 GBgood
Q5_K_M1537.2 GB1466.8 GBgood
Q6_K1844.5 GB1760.2 GBexcellent
Q8_02459.2 GB2346.9 GBexcellent
F164917.9 GB4693.9 GBlossless

How to run Qwen 3.8 2.4T-A95B 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 3.8 2.4T-A95B on Multi-GPU or High-RAM Desktop

  1. Plan for 1341.1 GB minimum combined memory and 2235.2 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 3.8 2.4T-A95B 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 1229.8 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 3.8 2.4T-A95B vs alternatives

ModelParamsContextMin RAMLicense
Qwen 3.8 2.4T-A95B2.4T (95B active)1M1341.1 GBQwen
Qwen 3 Coder 480B480B (35B active)256K268.2 GBApache 2.0
Qwen 3.5 397B-A17B397B (17B active)256K221.8 GBApache 2.0
Qwen 3 235B-A22B235B (22B active)128K131.3 GBApache 2.0
Qwen 3 VL 235B-A22B235B (22B active)256K131.3 GBApache 2.0
Qwen 3.8 Flash 128B125B (6B active)128K32 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 3.8 2.4T-A95B?
Qwen 3.8 2.4T-A95B is a 2.4T sparse mixture-of-experts model (95B active active per token) from Alibaba, supporting up to 1M tokens with 1341.1 GB minimum RAM and 2235.2 GB recommended RAM.
How do I download and install Qwen 3.8 2.4t?
Download a GGUF build for Qwen 3.8 2.4T-A95B, start with Q4_K_M, and confirm the model fits locally before increasing context or quant quality.
How much VRAM or RAM does Qwen 3.8 2.4T-A95B require?
At Q4_K_M, Qwen 3.8 2.4T-A95B occupies roughly 1229.8 GB of active memory and 1173.5 GB on disk. Add 2–6 GB of spare headroom for KV cache buffers, runtime scratch space, and OS apps (2235.2 GB recommended).
What is the best way to run Qwen 3.8 2.4T-A95B locally?
Run Qwen 3.8 2.4T-A95B in llama.cpp starting with Q4_K_M. Keep initial context at 4K–8K tokens and confirm 100% GPU/Metal offload before scaling context length.
Can Qwen 3.8 2.4T-A95B run on a consumer laptop or smartphone?
No smartphone or 8–16 GB laptop can hold Qwen 3.8 2.4T-A95B cleanly in local memory. It targets multi-GPU NVIDIA workstations, high-RAM Mac Studios, or MoE CPU-RAM hybrid rigs with at least 1341.1 GB total memory.
How does Qwen 3.8 2.4T-A95B compare to Qwen 3 Coder 480B?
Qwen 3 Coder 480B (480B, 268.2 GB min RAM) sits closest in size. Compare active parameters, context ceiling (1M vs 256K), and license terms (Qwen) before switching.