OpenWeightsTerminal
Pairs

Qwen3.8-Flash-Next

RTX 4060 Ti 16GB · IQ3_S · Strata

Decode · tok/s
55
Prefill
—
tok/100W
33.3
太神奇了,Strata升级到v0.1.44.1版本,我的4060ti 16g都能稳定平均55 tok/s ,感谢 @veit06 带来如此优秀的项目,让我们低配置也能玩这种级别的大模型 。Qwen3.8 Flash-Next IQ3_S . 262K一次成功完成率很高

First-person 4060 Ti 16GB on Strata v0.1.44.1. Rank is the quoted stable average 55. Prior pair from same user was 50 on unspecified VRAM size.

Officialclaimed
Open pair