Qwen3 235B A22B (Non-reasoning)
AlibabaQwenOpen WeightApache 2.0 · Commercial OK
Descripción
Qwen3 235B A22B is a large language model developed by Alibaba, featuring a Mixture-of-Experts (MoE) architecture with 235 billion total parameters and 22 billion activated parameters. It achieves competitive results in benchmark evaluations of coding, math, general capabilities, and more, compared to other top-tier models.
Fecha de lanzamiento
2025-04-28
Parámetros
235.0B
Longitud del contexto
131K
Modalidades
text
Radar de capacidades
33
general
23
coding
40
reasoning
39
scienceest.
70
agents
0
multimodal
Science usa un proxy de razonamiento cuando los benchmarks científicos dedicados no están disponibles.
Rankings
| Dominio | #Posición | Puntuación | Fuente |
|---|---|---|---|
| Code Ranking | 351 | 19.0 | AA |
| General Ranking | 286 | 38.0 | AA |
| Math Reasoning | 227 | 39.0 | AA |
| Reasoning | 32 | 79.0 | LS |
| Science | 275 | 40.0 | AA |
Puntuaciones de benchmarks (LLM Stats)
Biology
GPQA
47.5%Aut.
Chemistry
SuperGPQA
44.1%Aut.
Code
EvalPlus
0.78 / 100Aut.
LiveCodeBench
70.7%Aut.
Aider
61.8%Aut.
Creativity
Arena Hard
95.6%Aut.
Finance
MMLU
87.8%Aut.
MMLU-Pro
68.2%Aut.
General
MMLU-Redux
87.4%Aut.
MMMLU
86.7%Aut.
MBPP
0.81 / 100Aut.
LiveBench
77.1%Aut.
Include
73.5%Aut.
MultiLF
71.9%Aut.
BFCL
70.8%Aut.
MultiPL-E
65.9%Aut.
Language
BBH
88.9%Aut.
Math
GSM8k
94.4%Aut.
AIME 2024
85.7%Aut.
MGSM
83.5%Aut.
AIME 2025
81.5%Aut.
MATH
71.8%Aut.
Reasoning
CRUX-O
0.79 / 100Aut.
Índices de evaluación AA
Math Index23.7
Intelligence Index17.0
Coding Index14.0
Math 5000.9
Mmlu Pro0.8
Gpqa0.6
Ifbench0.4
Livecodebench0.3
Aime0.3
Scicode0.3
Tau20.3
Aime 250.2
Terminalbench Hard0.1
Hle0.0
Lcr0.0
Puntuaciones por categoría LLM Stats
Writing100
Creativity100
Language80
Math80
Reasoning80
Tool Calling70
Code70
Finance70
General70
Healthcare70
Legal70
Biology50
Chemistry50
Physics50
Economics40
Precios
Precio de entrada$0.45 / 1M tokens
Precio de salida$1.8 / 1M tokens
Precio mixto (3:1)$0.787 / 1M tokens
Velocidad
Tokens/seg64.1 tokens/s
Retraso del primer token1.24s
Tiempo hasta la respuesta1.24s
Proveedores disponibles
(Unidades internas LS)No hay datos de proveedores disponibles