Unverified50% confidenceBenchmarkExact time
Qwen 3.8在M5 Max上基础速度为22 token/s，开启MTP后可提升到34甚至56 token/s
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/10/2026
First Seen
Valid until: 12/9/2026
Sources
Related Entities
Related Claims
UnverifiedQwen3.7 Max已支持百万Token级别的上下文长度70% similarUnverified在Gemma 4 12B模型的Token生成测试中，Spark达26.4 tokens/s、Halo 24.6 tokens/s、M4 Pro最快33.8 tokens/s70% similarUnverified在Mac M3 Ultra上运行Qwen3系列量化模型，Q8量化下推理速度达到约34 tokens/s69% similarUnverifiedQwen2.5-Max-0902最大输出为13万Token，推理模式下上限可达26万Token68% similarUnverifiedQwen3.6 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%67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/887804API
curl https://kongchang.com/api/v1/knowledge/claims/887804MCP
get_claim(id=887804)