The letter

One letter a week, starting soon.

The week's signal, what shipped, what is worth running tonight, and the editor's note. No tracking, no ads, nothing else.

We keep your address, your language and the date you joined, nothing else. Every letter has a one-click unsubscribe link that deletes the record.

← Accept All   Archive
NVIDIA Technical BlogSilicon

ReplaySSM: Rethinking State Storage for Faster Hybrid-Model Decoding

10.12

Hybrid models interleave state-space model (SSM) layers with a few attention layers. NVIDIA Nemotron 3, Qwen3.5 through Qwen3.8, Kimi-K3, and GLM-5.3-Flash are...

Models & releasesAgentic AI / Generative AIAI InferencefeaturedInference Performance
Read at NVIDIA Technical Blog ↗

Related

More from NVIDIA Technical Blog on Accept All.