LaunchAI ModelsJuly 28, 2026
Liquid AI launches LFM2.5 encoders for fast long-context CPU inference
LFM2.5 encoders are designed to enable efficient long-context processing on CPUs. No specific benchmark or parameter counts provided.
1 source
More stories today
- GM triples merged pull requests with AI agent workflows
- Runway turns AI video avatar drift bug into feature
- NVIDIA Jetson AI Lab shows local GenAI with Ollama, vLLM, llama.cpp
- AI agents fail silently with no error logs or alerts
- Discovering cryptographic weaknesses with Claude