Qwen

Qwen3.5-122B-A10B

by Qwentext+image+video->text5 endpoints
Neura Intelligence Index
69.0/ 100
Rank
#88
Confidence
high
Benchmarks
9(23% coverage)

Overview

The Qwen3.5 122B-A10B native vision-language model is built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency. In terms of...

Capabilities

Text generation
Image understanding
Video understanding
Tool use / Function calling
Structured output
Extended reasoning
Long context
Top-P sampling
Stop sequences
Deterministic seed
Log probabilities

Modalities

Input
TextImageVideo
Output
Text

Technical Specifications

Context Window
262.1K tokens
Max Output
65.5K tokens
Tokenizer
Qwen3
Uptime (24h)
9993.1%
Parameters
122000000000B
License
apache_2_0Open Weights
Country
🇨🇳CN
Release Date
February 24, 20266mo

Supported Parameters

frequency_penaltyinclude_reasoninglogit_biaslogprobsmax_tokensmin_ppresence_penaltyreasoningrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_ktop_logprobstop_p

Benchmark Performance

general

BenchmarkScoreSource
MMMLU
86.7
verified

coding

BenchmarkScoreSource
SWE-bench Verified
72
verified

reasoning

BenchmarkScoreSource
GPQA Diamond
86.6
verified
Humanity's Last Exam
47.5
verified

multimodal

BenchmarkScoreSource
MMMU
83.9
verified
CharXiv Reasoning
77.2
verified
MMMU-Pro
76.9
verified
ScreenSpot Pro
70.4
verified

agent

BenchmarkScoreSource
BrowseComp
63.8
verified

Performance Over Time

Pricing

Per 1M tokens
Input$0.26
Output$2.08
Added
February 25, 2026
Last synced 9/16/2026