Saltar al contenido principal
AITier

Qwen3.8 Max (0902)

AlibabaQwenOpen WeightQwen3.8-Max License · Uso Comercial

Descripción

Qwen3.8 Max, released as the Qwen3.8-2.4T-A95B open-weight checkpoint, is Qwen's flagship mixture-of-experts model for coding, research, professional workflows, and long-horizon agents. It has 2.4 trillion total parameters with 95 billion active parameters, a native 262,144-token context window extensible to about 1 million tokens, and always-on thinking. QwenCloud serves the same model in the Qwen3.8 Max API family with text and image input.

Fecha de lanzamiento
2026-09-02
Parámetros
2.4T
Longitud del contexto
1.0M
Modalidades
image, pdf, text, video

Radar de capacidades

45
general
72
coding
93
reasoning
69
science
60
agents
90
multimodal

Rankings

Dominio#PosiciónPuntuaciónFuente
Capacidad agéntica18
61.0
LS
Ranking de codificación42
92.0
AA
Ranking general35
78.0
AA
Razonamiento matemático31
91.0
LB
Ranking multimodal1
84.0
LS
Razonamiento24
88.0
LB
Ciencia61
80.0
AA

Puntuaciones de benchmarks (LLM Stats)

(LLM Stats (zeroeval))

Agents

QwenSVG1713.00 / 2000Aut.
AndroidWorld85.3%Aut.
MobileWorld77.8%Aut.
AndroidBench75.1%Aut.
CoWorkBench74.8%Aut.
Toolathlon72.5%Aut.
Workspace Bench67.7%Aut.
Job Bench53.4%Aut.
OneMillion Bench52.5%Aut.
Agents' Last Exam52.4%Aut.
MLS-Bench Lite41.0%Aut.
AutomationBench27.3%Aut.

Code

QwenReactBench1724.00 / 2000Aut.
PaperBench93.0%Aut.
QwenSWEBench80.7%Aut.
FrontierSWE73.5%Aut.
SkillsBench70.2%Aut.
Vision2Web69.0%Aut.
QwenQoderBench58.4%Aut.
DeepSWE 1.156.6%Aut.
NL2Repo55.9%Aut.

Healthcare

HealthBench60.2%Aut.

Instruction Following

IFBench82.8%Aut.

Knowledge

PLawBench73.2%Aut.
PRBench-Finance58.3%Aut.
PRBench-Legal57.6%Aut.

Long Context

MRCR v2 (8-needle)92.9%Aut.
LongBench v266.3%Aut.

Multimodal

OSWorld-Verified86.1%Aut.

Reasoning

GPQANYU + Cohere + Anthropic (2023)92.6%Aut.
Terminal-Bench 2.186.6%Aut.
SWE-Bench ProPrinceton NLP (2024)67.7%Aut.
Humanity's Last Exam (with tools, text-only)56.2%Aut.
Humanity's Last Exam43.6%Aut.

Search

WideSearch81.9%Aut.

Vision

VideoMME w sub.90.4%Aut.
RealWorldQA88.0%Aut.
ScreenSpot Pro84.5%Aut.
MMMU-Pro82.3%Aut.
LVBench81.8%Aut.
ERQA77.8%Aut.
PerceptionBench63.5%Aut.

Índices de evaluación AA

(Artificial Analysis)
Gpqa(NYU + Cohere + Anthropic (2023))
92.8
Terminalbench V2 1
88.8
Lcr(Artificial Analysis)
80.3
Coding Index(Artificial Analysis)
76.2
Scicode(UIUC + Argonne National Lab (2024))
52.1
Tau Banking
47.8
Intelligence Index(Artificial Analysis)
45.4
Hle(Center for AI Safety + Scale AI (2025))
43.1
Terminalbench V4 0
38.9

Puntuaciones por categoría LLM Stats

(LLM Stats (zeroeval))
Physics
90
Biology
90
Chemistry
90
Video
90
Instruction Following
80
Multimodal
80
Search
80
Spatial Reasoning
80
Grounding
80
Vision
80
Long Context
70
Reasoning
70
Structured Output
70
General
70
Agents
70
Code
70
Productivity
60
Healthcare
60
Tool Calling
60
Math
50

Precios

Precio de entrada$2 / 1M tokens
Precio de salida$6 / 1M tokens
Precio mixto (3:1)$3 / 1M tokens
Precio de lectura caché$0.25 / 1M tokens
Precio de escritura caché$2.5 / 1M tokens

Velocidad

Tokens/seg40.0
Retraso del primer token1.84s
Tiempo hasta la respuesta51.88s

Ranking de Precios por Proveedor

Ranking de Precios por Proveedor

6 proveedores

Más barato: DeepInfraMás caro: EmpirioLabs AI
ProveedorEntradaSalida
1DeepInfraMás barato
$0
$0
2Novita
$0
$0.00001
3Fireworks
$0
$0.00001
4Together
$0
$0.00001
5AlibabaPRINCIPAL
$2
$6
6EmpirioLabs AI
$2
$6

Comparar precios entre diferentes proveedores de API para este modelo.

Fuentes externas

LLM StatsArtificial Analysis