मुख्य सामग्री पर जाएं
AITier

Qwen3.5 27B (Reasoning)

AlibabaQwenOpen WeightApache 2.0 · Commercial OK

विवरण

Qwen3.5-27B is a multimodal dense foundation model with 27 billion parameters. It combines strong reasoning, coding, multilingual, long-context, and visual understanding performance in a production-friendly open-weight package with a native 262K context window.

रिलीज़ तिथि
2026-02-24
पैरामीटर
27.0B
संदर्भ लंबाई
262K
मोडैलिटीज़
image, text, video

क्षमता रडार

38
general
36
coding
86
reasoning
57
scienceअनुमानित
60
agents
80
multimodal

समर्पित विज्ञान बेंचमार्क उपलब्ध न होने पर Science तर्क प्रॉक्सी का उपयोग करके अनुमान लगाता है।

रैंकिंग

डोमेन#रैंकस्कोरस्रोत
Agents & Tools51
57.0
LS
Code Ranking84
65.0
AA
General Ranking43
80.0
AA
Multimodal Ranking60
70.0
LS
Reasoning54
67.0
LS
Science63
68.0
AA

बेंचमार्क स्कोर (LLM Stats)

3d

SUNRGBD0.35 / 100स्वयं
Hypersim0.13 / 100स्वयं

Agents

t2-bench79.0%स्वयं
BFCL-V468.5%स्वयं
AndroidWorld_SR64.2%स्वयं
WideSearch61.1%स्वयं
BrowseComp61.0%स्वयं
FullStackBench en60.1%स्वयं
TIR-Bench59.8%स्वयं
FullStackBench zh57.4%स्वयं
OSWorld-Verified56.2%स्वयं
VITA-Bench41.9%स्वयं
Terminal-Bench 2.041.6%स्वयं
DeepPlanning22.6%स्वयं

Biology

GPQA85.5%स्वयं

Chemistry

SuperGPQA65.6%स्वयं

Code

SWE-Bench Verified72.4%स्वयं

Communication

Multi-Challenge60.8%स्वयं

Embodied

EmbSpatialBench0.84 / 100स्वयं

Finance

MMLU-Pro86.1%स्वयं
MMLU-ProX82.2%स्वयं

General

IFEval95.0%स्वयं
MMLU-Redux93.2%स्वयं
C-Eval90.5%स्वयं
MAXIFE88.0%स्वयं
Global PIQA87.5%स्वयं
MMMLU85.9%स्वयं
MMMU82.3%स्वयं
Include81.6%स्वयं
MMStar81.0%स्वयं
LiveCodeBench v680.7%स्वयं
IFBench76.5%स्वयं
MMMU-Pro75.0%स्वयं
LongBench v260.6%स्वयं
NOVA-6358.1%स्वयं
SimpleVQA0.56 / 100स्वयं

Grounding

RefCOCO-avg0.91 / 100स्वयं
ScreenSpot Pro70.3%स्वयं
RefSpatialBench0.68 / 100स्वयं

Healthcare

VideoMMMU82.3%स्वयं
SlakeVQA80.0%स्वयं
MedXpertQA62.4%स्वयं
PMC-VQA62.4%स्वयं

Image To Text

OCRBench89.4%स्वयं

Language

LingoQA82.0%स्वयं
WMT24++77.6%स्वयं

Long Context

MLVU85.9%स्वयं
LVBench73.6%स्वयं
AA-LCR66.1%स्वयं
MMLongBench-Doc0.60 / 100स्वयं

Math

HMMT 202592.0%स्वयं
HMMT2589.8%स्वयं
MathVista-Mini87.8%स्वयं
DynaMath87.7%स्वयं
MathVision86.0%स्वयं
CodeForces0.81 / 3000स्वयं
PolyMATH71.2%स्वयं
Humanity's Last Exam48.5%स्वयं

Multimodal

VLMsAreBlind96.9%स्वयं
V*93.7%स्वयं
AI2D92.9%स्वयं
MMBench-V1.192.6%स्वयं
OmniDocBench 1.588.9%स्वयं
VideoMME w sub.87.0%स्वयं
VideoMME w/o sub.82.8%स्वयं
CC-OCR81.0%स्वयं
CharXiv-R79.5%स्वयं
MVBench74.6%स्वयं
MMVU73.3%स्वयं
BabyVision44.6%स्वयं
ZEROBench-Sub0.36 / 100स्वयं
Nuscene15.2%स्वयं
ZEROBench0.10 / 100स्वयं

Reasoning

CountBench0.98 / 100स्वयं
Hallusion Bench70.0%स्वयं
BrowseComp-zh62.1%स्वयं
ERQA60.5%स्वयं
Seal-047.2%स्वयं
OJBench40.1%स्वयं

Spatial Reasoning

RealWorldQA83.7%स्वयं

Vision

ODinW41.1%स्वयं

AA मूल्यांकन सूचकांक

Intelligence Index
42.1
Coding Index
34.9
Tau2
0.9
Gpqa
0.9
Ifbench
0.8
Lcr
0.7
Scicode
0.4
Terminalbench Hard
0.3
Hle
0.2

LLM Stats श्रेणी स्कोर

Biology
90
Instruction Following
90
Structured Output
80
Text-to-image
80
Video
80
Chemistry
80
Embodied
80
Finance
80
General
80
Grounding
80
Image To Text
80
Language
80
Legal
80
Math
80
Physics
80
Spatial Reasoning
70
Vision
70
Economics
70
Frontend Development
70
Healthcare
70
Long Context
70
Multimodal
70
Reasoning
70
Tool Calling
60
Agents
60
Code
60
Communication
60
Search
60
Spatial
20
3d
20

मूल्य निर्धारण

इनपुट मूल्य$0.3 / 1M tokens
आउटपुट मूल्य$2.4 / 1M tokens
मिश्रित मूल्य (3:1)$0.825 / 1M tokens

गति

टोकन/सेकंड87.6 tokens/s
पहले टोकन में देरी1.40s
पहले उत्तर में देरी24.23s

उपलब्ध प्रदाता

(LS आंतरिक इकाइयाँ)
प्रदाताइनपुट मूल्यआउटपुट मूल्य
Novita300K2.4M

बाहरी लिंक

LLM Stats