Product
StraightA
An open-source local large-model inference engine that uses quantization and memory offloading techniques to enable consumer-grade GPUs with limited VRAM to run extremely large parameter models. Released under the MIT license.
Timeline (last 90 days)
Oct 8
The open-source project StraightA claims to be able to run the 125B-parameter Qwen 3.8 Flash Next model on a consumer-grade PC with more than 12G of VRAM, provided the system is equipped with 35 to 55G of RAM.
Unverified50%
Oct 8
The StraightA project is open-sourced under the MIT license.
Unverified50%