Expired50% confidenceFactExact time
LFM2.5 has 8.3 billion total parameters but only 1.5 billion parameters activated per inference
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
Valid until: 9/30/2026(expired)
Sources
LFM2.5 Local Deployment Hands-On: An 8B Parameter Model That Outperforms GPT-o3s in Tool Calling
bilibiliAGI_Ananas6/1/2026
Related Claims
UnverifiedLFM2.5-8B-A1B总参数量为8B，推理时仅激活1.5B参数79% similarVerifiedGLM-4.6 adopts a Mixture of Experts (MoE) architecture, with a total parameter count of 355 billion, activating approximately 32 billion parameters during inference.75% similarUnverifiedGLM5 has approximately 745B parameters, with approximately 44B activated parameters, giving a parameter activation ratio of approximately 1:17.70% similarUnverifiedHy3 adopts a Mixture of Experts (MoE) architecture with a total parameter count of 295 billion (295B), activating only 21 billion (21B) parameters per inference.66% similarUnverifiedLFM2.5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66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/47920API
curl https://kongchang.com/api/v1/knowledge/claims/47920MCP
get_claim(id=47920)