Unverified50% confidenceBenchmarkExact time
V4.1-Flash在某编程测试维度的得分达到74.2分,相比V4-Flash提升了将近20分
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/11/2026
First Seen
Valid until: 12/10/2026
Sources
Related Entities
Related Claims
UnverifiedV4.1 Flash提供低、中、最高三档推理强度,八项推理基准平均分从67.1涨到76.3,代价是输出Token增加约2.5倍77% similarUnverified盘古2.0 Flash在可验证指令遵循测试中得分95.9,位列全场第一73% similarUnverified盘古2.0 Flash在AIME类数学测试中得分93.3,在HMMT测试中得分91.5,数学能力处于第一梯队70% similarUnverifiedNVFP4量化方案相比BF16基线实现了1.5倍的推理加速,同时在top-1准确率上保留了92%到97%的水平67% similarUnverified在综合智能指数AA上,Qwen3.8 Flash Next得分56,Qwen3.8 Flash云端版得分57,DeepSeek V4 Flash得分5265% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/900378API
curl https://kongchang.com/api/v1/knowledge/claims/900378MCP
get_claim(id=900378)