Unverified50% confidenceBenchmarkExact time
Mistral 3.8B 指令模型使用 Q4 量化和 llama.cpp 原生调用时，综合 Score 从 27.3% 提升到 78.3%，区别仅在于是否有 Forge 护栏
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/18/2026
First Seen
Valid until: 10/16/2026
Sources
Related Claims
Unverified原版Qwen3.6 27B Q4在工具调用测试中获得97分，说明蒸馏并不总能提升性能67% similarUnverifiedQwen3.5的参数激活率为4.3%，而DeepSeek V3的671B模型激活率为5.5%66% similarUnverifiedOmniMerge V4在MBPP代码测试中达到73.4%，相比原版Q6K量化的56.2%提升了15.8个百分点66% similarUnverified在终端基准（Terminal Bench）测试中，Qwen 3.6 27B得分约60.7，而Muse Glimmer只有51.7，明显落后65% similarUnverifiedQwen 3.6 Max在中性模式下端到端准确率仅12.3%64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/547700API
curl https://kongchang.com/api/v1/knowledge/claims/547700MCP
get_claim(id=547700)