Alibaba

Qwen 3.5 122B A10B

Name: Qwen 3.5 122B A10B
Rating: 93 (29 reviews)
Author: Alibaba

Frontier

HuggingFace

Jun 2025Veröffentlicht131K TokenKontextApache 2.0Lizenz98 HerausragendQualität

Qwen 3.5 122B A10B (122B parameters) requires approximately 78.4 GB of VRAM with Q4_K_M quantization. As a Mixture of Experts model with 10B active parameters, it uses less memory than its total parameter count suggests. For the best balance of quality and speed, we recommend hardware with at least 91 GB of VRAM.

Loslegen

— kopieren & einfügen, um lokal auszuführen

Copy-paste commands to run Qwen 3.5 122B A10B on your machine.

Run

lms load Qwen3.5-122B-A10B-Instruct && lms server start

Quick specs

Parameters122B (10B active)

Architecturemoe (MoE)

Context131K tokens

Modalitytext

Min RAM47.6 GB

Rec. RAM74.4 GB (Q4_K_M)

LicenseApache 2.0

FamilyQwen

✓ Code✓ Chat✓ Reasoning✓ RAG

About this model

Qwen3.5 122B A10B is a high-quality MoE model with 10B active parameters. Strong performance across coding, reasoning, and RAG with moderate inference cost.

•122B total params with 10B active — excellent quality-to-cost ratio
•Competitive with dense 32-70B models on reasoning benchmarks
•Requires 48+ GB VRAM at Q4_K_M — suitable for A6000, dual GPU setups, or M-series Macs

Verwandte Modelle

Schnellauswahl

Bestes BudgetS

Mac Studio M3 Ultra 256GB~$6,999 — 35 tok/s

Beste GesamtwahlS

AMD Instinct MI250X 128GB~$15,000 — 100 tok/s

Beste Hardware

Top-Empfehlungen für Qwen 3.5 122B A10B

AMD Instinct MI250X 128GBS

128 GB

AMD Instinct MI300A 128GBS

128 GB

Gaudi 3 128GBS

128 GB

AMD Instinct MI250 128GBS

128 GB

Intel Data Center GPU Max 1550 128GBS

128 GB

Dieses Modell ausführen

Qwen 3.5 122B A10B on AMD Instinct MI250X 128GB Qwen 3.5 122B A10B on AMD Instinct MI300A 128GB Qwen 3.5 122B A10B on Gaudi 3 128GB

Quantisierungsoptionen

VRAM-Schätzungen nach Quantisierungsstufe

No hardware detected — fit column shows raw VRAM estimates

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	47.6 GB	Low	—
Q3_K_S	3	59.8 GB	Low	—
NVFP4	4	68.3 GB	Medium	—
Q4_K_M	4	74.4 GB	Medium	—
Q5_K_M	5	87.8 GB	High	—
Q6_K	6	100.0 GB	High	—
Q8_0	8	130.5 GB	Very High	—
F16	16	250.1 GB	Maximum	—

Quality benchmarks

Qwen 3.5 122B A10B benchmark scores

Benchmark verified

Coding

SWE-bench Verified72.0%

HumanEval+—

Aider Polyglot—

LiveCodeBench78.9%

Reasoning

MMLU-Pro86.1%

GPQA Diamond85.5%

MATH-500—

ARC Challenge—

General

Chatbot Arena—

IFEval93.4%

Source: official · 2025-06-25

Hardware-Kompatibilität

Eignungsschätzungen für alle Hardware

Rechner öffnen

Computing compatibility...

Speicheraufschlüsselung

Reference: RTX 2060 6GB

Weights74.4 GB

KV Cache2.4 GB

Runtime0.9 GB

Headroom0.6 GB

Häufig gestellte Fragen

FAQ — Qwen 3.5 122B A10B

Siehe auch

Quantisierungsleitfaden Bewertungsmethodik Rechner öffnen