Qwen3.5 122B A10B (Non-reasoning)

AlibabaQwen開源權重Apache 2.0 · 商用許可

描述

Qwen3.5-122B-A10B is a multimodal Mixture-of-Experts model with 122 billion total parameters and 10 billion activated parameters. It combines strong reasoning, coding, long-context, and visual understanding performance with production-friendly efficiency and a native 262K context window.

發布日期

2026-02-24

參數規模

122.0B

上下文長度

262K

支援模態

audio, image, text, video

能力雷達圖

general

coding

reasoning

science

agents

multimodal

排行榜排名

領域	#排名	分數	來源
智慧體能力模型榜	79	45.0	LS
程式碼能力榜	184	57.0	AA
通用能力榜	193	57.0	AA
多模態榜	24	59.0	LS
科學能力	181	57.0	AA

基準測試分數 (LLM Stats)

(LLM Stats (zeroeval))

3d

SUNRGBD

0.36 / 100自報

Hypersim

0.13 / 100自報

Agents

t2-bench

79.5%自報

BFCL-V4

72.2%自報

AndroidWorld_SR

66.4%自報

BrowseCompOpenAI (2025)

63.8%自報

FullStackBench en

62.6%自報

WideSearch

60.5%自報

FullStackBench zh

58.7%自報

OSWorld-Verified

58.0%自報

TIR-Bench

53.2%自報

Terminal-Bench 2.0Stanford × Laude Institute (2026)

49.4%自報

VITA-Bench

33.6%自報

DeepPlanning

24.1%自報

Biology

GPQANYU + Cohere + Anthropic (2023)

86.6%自報

Chemistry

SuperGPQA

67.1%自報

Code

SWE-Bench Verified

72.0%自報

Communication

Multi-Challenge

61.5%自報

Embodied

EmbSpatialBench

0.84 / 100自報

Finance

MMLU-Pro

86.7%自報

MMLU-ProX

82.2%自報

General

MMLU-Redux

94.0%自報

IFEvalGoogle Research (2023)

93.4%自報

C-Eval

91.9%自報

Global PIQA

88.4%自報

MAXIFE

87.9%自報

MMMLU

86.7%自報

MMMU

83.9%自報

MMStar

82.9%自報

Include

82.8%自報

LiveCodeBench v6

78.9%自報

MMMU-Pro

76.9%自報

IFBench

76.1%自報

SimpleVQA

0.62 / 100自報

LongBench v2

60.2%自報

NOVA-63

58.6%自報

Grounding

RefCOCO-avg

0.91 / 100自報

ScreenSpot Pro

70.4%自報

RefSpatialBench

0.69 / 100自報

Healthcare

VideoMMMU

82.0%自報

SlakeVQA

81.6%自報

MedXpertQA

67.3%自報

PMC-VQA

63.3%自報

Image To Text

OCRBench

92.1%自報

Language

LingoQA

80.8%自報

WMT24++

78.3%自報

Long Context

MLVU

87.3%自報

LVBench

74.4%自報

AA-LCR

66.9%自報

MMLongBench-Doc

0.59 / 100自報

Math

HMMT 2025

91.4%自報

HMMT25

90.3%自報

MathVista-Mini

87.4%自報

MathVision

86.2%自報

DynaMath

85.9%自報

CodeForces

0.85 / 3000自報

PolyMATH

68.9%自報

Humanity's Last Exam

47.5%自報

Multimodal

VLMsAreBlind

96.7%自報

AI2D

93.3%自報

93.2%自報

MMBench-V1.1

92.8%自報

OmniDocBench 1.5

89.8%自報

VideoMME w sub.

87.3%自報

VideoMME w/o sub.

83.9%自報

CC-OCR

81.8%自報

CharXiv-R

77.2%自報

MVBench

76.6%自報

MMVU

74.7%自報

BabyVision

40.2%自報

ZEROBench-Sub

0.36 / 100自報

Nuscene

15.4%自報

ZEROBench

0.09 / 100自報

Reasoning

CountBench

0.97 / 100自報

BrowseComp-zh

69.9%自報

Hallusion Bench

67.6%自報

ERQA

62.0%自報

Seal-0

44.1%自報

OJBench

39.5%自報

Spatial Reasoning

RealWorldQA

85.1%自報

Vision

ODinW

44.5%自報

AA 評測指數

(Artificial Analysis)

Coding Index(Artificial Analysis)

43.3

Intelligence Index(Artificial Analysis)

28.2

Tau2(Sierra + U Toronto + Vector Institute (2025))

0.8

Gpqa(NYU + Cohere + Anthropic (2023))

0.8

Lcr(Artificial Analysis)

0.6

Ifbench(Google Research (2023))

0.5

Terminalbench V2 1

0.5

Scicode(UIUC + Argonne National Lab (2024))

0.4

Terminalbench Hard(Stanford × Laude Institute (2026))

0.3

Hle(Center for AI Safety + Scale AI (2025))

0.2

Tau Banking

0.1

LLM Stats 分類評分

(LLM Stats (zeroeval))

Biology

Legal

Math

Physics

Structured Output

Image To Text

Instruction Following

Language

Embodied

Finance

Grounding

Healthcare

Chemistry

Text-to-image

Video

Long Context

Multimodal

Reasoning

Spatial Reasoning

Frontend Development

General

Economics

Vision

Agents

Code

Communication

Tool Calling

Spatial

定價

輸入價格$0.4 / 1M tokens

輸出價格$3.2 / 1M tokens

混合價格(3:1)$1.1 / 1M tokens

速度

Tokens/秒148.9

首Token延遲1.03s

首回答延遲1.03s

供應商價格排行

3 個供應商

最便宜: Alibaba最貴: Cortecs

供應商輸入輸出

1Alibaba主要

$0.4

$3.2

2NanoGPT

$0.437

$3.496

3Cortecs

$0.495

$3.46

比較該模型在不同 API 供應商之間的定價。

外部連結

LLM Stats Artificial Analysis