# Researchers from Stanford and UC Berkeley Introduce TokenGate: Token-Gated Attention Makes Long-Context Prefill Up to 11.3x Faster

> **Publication**: aitalky (Independent AI Journalism)
> **Canonical Citation URL**: https://aitalky.vercel.app/news/researchers-from-stanford-and-uc-berkeley-introduce-tokengate-token-gated-attention
> **Original Source**: [MarkTechPost](https://www.marktechpost.com/2026/10/25/stanford-berkeley-tokengate-learned-token-pruning-llm-prefill/)
> **Author**: Asif Razzaq (undefined)
> **Published Date**: Sunday, October 25, 2026
> **Category**: RESEARCH
> **Tags**: AI, MarkTechPost, research

## Executive Summary
Stanford and UC Berkeley researchers released TokenGate, a drop-in learned token pruning module for LLMs. It cuts long-context prefill time-to-first-token by 1.4x to 11.3x and peak memory by 1.5x to 3.4x without accuracy loss



## Full Coverage
Stanford and UC Berkeley researchers released TokenGate, a drop-in learned token pruning module for LLMs. It cuts long-context prefill time-to-first-token by 1.4x to 11.3x and peak memory by 1.5x to 3.4x without accuracy loss

---
*Generated for AI agents and LLM citation by aitalky (https://aitalky.vercel.app).*