Will It Run AI

AlibabaAlibaba

Qwen 3 32B

Frontera
huggingfaceHuggingFaceollamaOllama
5.1MDescargas697Me gustaApr 2025Publicado131K tokensContextoApache 2.0Licencia95 ExcepcionalCalidad

Qwen 3 32B (32B parameters) requires approximately 25.2 GB of VRAM with Q4_K_M quantization. For the best balance of quality and speed, we recommend hardware with at least 29 GB of VRAM.

Comenzar

— copia y pega para ejecutar en local

Copy-paste commands to run Qwen 3 32B on your machine.

Run

ollama run qwen3:32b

Quick specs

Parameters32B
Architecturedense
Context131K tokens
Modalitytext
Min RAM12.5 GB
Rec. RAM19.5 GB (Q4_K_M)
LicenseApache 2.0
FamilyQwen
✓ Chat✓ Reasoning

About this model

Qwen3 es la última generación de grandes modelos de lenguaje de la serie Qwen, que ofrece una suite completa de modelos densos y de mezcla de expertos (MoE). Construido con un entrenamiento extensivo, Qwen3 logra avances significativos en razonamiento, seguimiento de instrucciones, capacidades agénticas y soporte multilingüe.

  • •Uniquely support of seamless switching between thinking mode: (for complex logical reasoning, math, and coding) and non-thinking mode (for...
  • •Significantly enhancement in its reasoning capabilities: , surpassing previous QwQ (in thinking mode) and Qwen2.5 instruct models (in non-thinking...
  • •Superior human preference alignment: , excelling in creative writing, role-playing, multi-turn dialogues, and instruction following, to deliver a...
  • •Expertise in agent capabilities: , enabling precise integration with external tools in both thinking and unthinking modes and achieving leading...
  • •Support of 100+ languages and dialects: with strong capabilities for multilingual instruction following and translation

Modelos relacionados

Tu hardware

Detectando...

Selecciones rápidas

Mejor económicoS
Mac mini M4 64GB~$1,099 — 9 tok/s
NVIDIA
Mejor en generalS
NVIDIA A100 40GB~$10,000 — 73 tok/s

Mejor hardware

Mejores opciones para Qwen 3 32B

NVIDIA
NVIDIA A100 40GBS
40 GB
NVIDIA
RTX PRO 5000 Blackwell 48GBS
48 GB
MacBook Pro M4 Max 64GBS
64 GB
NVIDIA
RTX 5090 32GBS
32 GB
NVIDIA
RTX 6000 Ada 48GBS
48 GB

Ejecutar este modelo

Qwen 3 32B on NVIDIA A100 40GBQwen 3 32B on RTX PRO 5000 Blackwell 48GBQwen 3 32B on MacBook Pro M4 Max 64GB

Opciones de cuantización

Estimaciones de VRAM por nivel de cuantización

No hardware detected — fit column shows raw VRAM estimates

QuantBitsVRAMQualityFit
Q2_K
2
12.5 GB
Low—
Q3_K_S
3
15.7 GB
Low—
NVFP4
4
17.9 GB
Medium—
Q4_K_M
4
19.5 GB
Medium—
Q5_K_M
5
23.0 GB
High—
Q6_K
6
26.2 GB
High—
Q8_0
8
34.2 GB
Very High—
F16
16
65.6 GB
Maximum—

Quality benchmarks

Qwen 3 32B benchmark scores

Benchmark verified

Coding

SWE-bench Verified—
HumanEval+—
Aider Polyglot—
LiveCodeBench65.7%

Reasoning

MMLU-Pro—
GPQA Diamond68.4%
MATH-50097.2%
ARC Challenge—

General

Chatbot Arena—
IFEval85.0%

Source: official · 2025-05-15

Compatibilidad de hardware

Estimaciones de encaje en todo el hardware

Abrir calculadora

Computing compatibility...

Desglose de memoria

Reference: RTX 2060 6GB

Weights19.5 GB
KV Cache3.9 GB
Runtime1.2 GB
Headroom0.6 GB

Preguntas frecuentes

FAQ — Qwen 3 32B

Ver también

Guía de cuantizaciónMetodología de puntuaciónAbrir calculadora