Unverified50% confidenceFactExact time

1-bit LLM 的核心思路是将模型权重压缩到每个参数仅约1比特的水平

1

Sources

50%

Confidence

Long-term

Relevance

7/17/2026

First Seen

Sources

Bonsai 27B：首个在手机运行的1-bit大模型，端侧AI新突破

Related Claims

Unverified1-bit量化（BitNet）将每个权重压缩为单个比特，仅表示+1或-176% similar Verified一个7B参数的模型经过4-bit量化后仅需约4GB显存即可运行73% similar VerifiedGGUF格式的4-bit量化可将模型权重从32位浮点压缩到4位整数，模型体积缩小约8倍72% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/544052

API

curl https://kongchang.com/api/v1/knowledge/claims/544052

MCP

get_claim(id=544052)