Unverified50% confidenceFactExact time
LFM2.5在BFCL-v3和BFCL-v4工具调用评测中分别比上代提升了19和23个点
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
Valid until: 9/30/2026
Sources
LFM2.5本地部署实测：8B参数碾压GPT-o3s的工具调用能力
bilibiliAGI_Ananas6/1/2026
Related Claims
UnverifiedLFM2.5在多项基准测试中展现出与体量高达其4倍模型相当的竞争力63% similarUnverified引导模式下平均端到端准确率从28.6%提升到33.7%，歧义检测F1值从45.3%提升到64.9%63% similarUnverifiedQwen3.5在HLE博士级综合题评测中得分28.7，而GPT-5.2为35.563% similarUnverifiedGlimmer智能指数35分，超越Gemma 2 27B（30分），几乎追平Qwen 2.5（36分），Qwen 2.5 72B以38分居榜首62% 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62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/49057API
curl https://kongchang.com/api/v1/knowledge/claims/49057MCP
get_claim(id=49057)