Expired60% confidenceFactExact time
LFM2.5 has 24 layers total, with 18 using Liquid AI's proprietary LIV (Liquid) convolution and only 6 retaining traditional attention mechanisms
2
Sources
60%
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 uses a Mixture of Experts (MOE) architecture combined with Liquid convolution69% similarUnverifiedLLaMA-2 70B采用80层、64个注意力头的架构，并使用GQA（8个KV头共享64个Q头）65% similarUnverifiedLiquid AI first publicly introduced liquid convolution technology in LFM1.065% similarUnverifiedGLM-5.3-Flash adopts a hybrid architecture of sparse attention and linear attention, and runs entirely on a domestically produced AI chip cluster.60% similarUnverifiedLLaMA-2 70B 模型共有 80 个 Transformer 层，每层包含自注意力机制和前馈神经网络两大模块59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/47923API
curl https://kongchang.com/api/v1/knowledge/claims/47923MCP
get_claim(id=47923)