VLM Performance：Qwen3.6 is natively multimodal, and Qwen3.6-35B-A3B showcases perception and multimo...
- Qwen3.6-35B-A3B 为原生多模态架构，非后期对齐
- 激活参数仅约30亿，但性能接近Claude Sonnet 4.5
- 在RefCOCO（92.0）和ODInW13（50.8）等空间理解任务表现突出
Qwen on X: "VLM Performance：Qwen3.6 is natively multimodal, and Qwen3.6-35B-A3B showcases perception and multimodal reasoning capabilities that far exceed what its size would suggest, with only around 3 billion activated parameters. Across most vision-language benchmarks, its performance https://t.co/nOVBNlVfzW" / X
Don’t miss what’s happening
People on X are the first to know.
Post
See new posts
Conversation
VLM Performance：Qwen3.6 is natively multimodal, and Qwen3.6-35B-A3B showcases perception and multimodal reasoning capabilities that far exceed what its size would suggest, with only around 3 billion activated parameters. Across most vision-language benchmarks, its performance matches Claude Sonnet 4.5, and even surpasses it on several tasks. Its strengths are particularly evident in spatial intelligence, where it achieves 92.0 on RefCOCO and 50.8 on ODInW13.
·
6
24
393
50
Read 6 replies
New to X?
Sign up now to get your own personalized timeline!
Sign up with Apple
By signing up, you agree to the Terms of Service and Privacy Policy, including Cookie Use.
Relevant people
-  Qwen @Alibaba_Qwen Follow Click to Follow Alibaba_Qwen Open foundation models for AGI.
Trending now
What’s happening
Sports · Trending
Jose Ramirez
Sports · Trending
Jaylen Brown
Only on X · Trending
Sunday Funday
Sports · Trending
Juan Brito
|
|
|
|
|
More
© 2026 X Corp.