Inkling-2-Mini hits 170 tok/s on 2x3090s in local demos

LiTianleli · x · 2026-11-21

loktar00 shares a local demo of Inkling-2-Mini, calling it a fantastic local model and reporting 170 tok/s on 2x RTX 3090s, with 4x3090 setups in progress and more benchmarks promised. Another retweet adds that the model outperforms Qwen3 8B and is much faster with the right inference engine.

Related event: Inkling 2 Mini Benchmarks: ~200 tok/s on a Single RTX 4090(5 posts)→

Original post →

More from Infra

Infra channel →