Unverified50% confidenceFactExact time
Qwen3.6 34B 80层模型在LM Studio上的MTP推测解码接受率约为78%，而原版27B仅约57-58%
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
Valid until: 9/30/2026
Sources
Qwen3.6非官方衍生模型：34B深度扩展与14B极致压缩解析
bilibiliAI观模者6/21/2026
Related Claims
UnverifiedQwen3.6 27B 在 SWE-bench Verified 上拿到77.2分73% similarUnverified开发者对Qwen3-VL-4B-Instruct执行abliteration处理后，HarmBench合规率达到100%，而原始模型仅为30.8%70% similarUnverifiedQwen3旗舰版（235B）的参数量仅为DeepSeek旗舰版（671B）的三分之一左右70% similarUnverified有社区网友猜测Qwen可能跳过35B规格直接推出122B级别的MoE模型69% similarUnverifiedQwen3.6-27B各项基准测试成绩在同规模（20B-30B参数级别）模型中处于领先地位69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/40888API
curl https://kongchang.com/api/v1/knowledge/claims/40888MCP
get_claim(id=40888)