Large Language ModelsQwen 3.8 Flash Next 125B Model can now Work at 100T/s Speed on Consumer GPU
Thanks to the Strata project, the 125 billion parameter Qwen model has been able to run at 100 tokens/second speed in consumer hardware such as RTX 40
#Qwen
6 articles · Posts with this tag
Large Language Models
Large Language ModelsKev: Compact Decision Models Built on Qwen3.5
Developed by Jared Palmer, Kev is a new family of light and efficient decision models formed on the basis of Qwen3.5.
Large Language ModelsQwen-Image-2.1 Offers New Solution to Create Compact and Efficient Image
Alibaba’s Qwen team introduced a new model with high quality image generation capacity using less welding
Large Language ModelsQwen 3.8 27B Offers Affordable Solution With Low VRAM Need
Qwen 3.8 27B model introduced by Shapelearn, only 13.1 Provides strong language processing capabilities with GB VRAM.
Large Language ModelsMore Smart Models with Less Data: Revolution in Artificial Intelligence Learning Paradigma
The researchers discovered that large language models can gain strong logical abilities while exposure to much less data instead of millions of tokens
Large Language ModelsRun Qwen 3.8 27B Model locally on Mac Studio: Real Performance Numbers
A developer has successfully operated the powerful Qwen 3.8 27B language model on Mac Studio and shared performance metrics.