Unverified50% confidenceFactExact time

9x compression means the storage and memory footprint of the model is reduced to approximately 11% of its original size.

1

Sources

50%

Confidence

Long-term

Relevance

9/17/2026

First Seen

Sources

Bonsai 2 27B: Exploring Near-Lossless Model Compression at 9x Smaller Footprint

Related Entities

Bonsai 2 27Bproduct

Related Claims

Verified4-bit量化理论上可将模型体积缩小8倍（相比32位存储）73% similar Unverifiedllama.cpp通过INT4/INT8量化将模型权重压缩至原始大小的25%以下73% similar Unverified8GB统一内存在运行较大参数量模型时相当捉襟见肘，留给操作系统和其他应用的空间极为有限72% similar Verified4-bit量化使模型存储空间缩减为FP16的约1/472% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/930399

API

curl https://kongchang.com/api/v1/knowledge/claims/930399

MCP

get_claim(id=930399)