Unverified50% confidenceBenchmarkExact time
DASA's experiments were validated on 6 models from the Llama and Qwen model families, covering parameter scales from 1B to 32B.
1
Sources
50%
Confidence
Long-term
Relevance
9/30/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe Qwen series of models is developed by Alibaba's Tongyi Lab, with 4B and 27B representing the parameter scales.71% similarUnverified研究表明当模型参数规模达到3B以上时，1.58-bit模型在多项基准测试中表现可接近甚至匹配同等规模的全精度模型69% similarUnverifiedBilibili developer Luke conducted a cross-evaluation test of the Qwen3 series 27B model released by Unsloth across multiple quantization levels from Q1 to Q8.68% similarUnverifiedModel selection must consider cost on the premise of acceptable performance, and evaluation can provide quantitative basis for choosing models of different scales such as 32B, 16B, 8B, and 2B67% similarUnverified万象模型提供了 1.3B 和 14B 参数量的不同规模版本66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/961004API
curl https://kongchang.com/api/v1/knowledge/claims/961004MCP
get_claim(id=961004)