MiniMax Cut Attention Compute by 28x at 1M Tokens

digitado ⋅ 19 de June de 2026

And Open-Sourced the Kernel

Like 0

Liked Liked

Search

Posts recentes

Train Your Own AI Image Detector: Why Off-the-Shelf Detectors Fail on Your Data (DINOv2 + ConvNeXt…
A 3B Model Just Matched DeepSeek V3.2 on Math (671B Parameters, 223x Larger)
From Chat to Cron: 11 Stages to a Self-Running Claude Assistant
MiniMax Cut Attention Compute by 28x at 1M Tokens
The CEO of Allbirds’ new AI biz has a plan, but no employees

No comments to show.