Unverified50% confidenceFactExact time
Bonsai compresses a 50-plus GB model down to 5–7GB while retaining considerable logical reasoning, coding, and multi-turn dialogue capabilities.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/10/2026
First Seen
Valid until: 1/8/2027
Sources
Related Entities
Related Claims
UnverifiedThe core claim of the Bonsai 2 27B project is to compress the model size to one-ninth of the original while keeping the model's capabilities nearly intact.71% similarUnverified1-bit Bonsai的智能密度得分为0.53/GB，是全精度模型的10倍以上，是最优低比特方案的约2.7倍69% similarUnverifiedBonsai 2 is a ternary-quantized large model released by Prism ML, retrained from scratch on a Qwen3-27B base, with an overall file size of approximately 6 GB.68% similarUnverifiedBonsai uses 1-bit quantization, representing each weight with just one bit, making the entire model file only slightly over 1GB.63% similarUnverifiedBonsai uses ternary quantization, compressing each weight to one of three values—-1, 0, or +1—with a shared scaling factor per 128 weights, requiring approximately 1.7 bits per weight, roughly 9 times smaller than 16-bit full precision.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/995013API
curl https://kongchang.com/api/v1/knowledge/claims/995013MCP
get_claim(id=995013)