LFM2.5-Encoders for Fast Long-Context Inference on CPU
Running long-context AI efficiently on standard computers could expand access beyond cloud infrastructure and reduce deployment costs.
Get smart on it
Two new encoder models have been released that can process long documents while running on standard computer processors. These models match the quality of larger systems on language understanding tasks while maintaining fast speeds even as input length increases to 8,192 tokens. The models are useful for production applications like text classification, intent routing, and safety filtering that need to run constantly and affordably on CPU hardware. The encoders were created by converting decoder models into bidirectional encoders through attention mask changes and training with masked language modeling, then extending their context length in a second training stage.
Related news
Pokee AI Releases Pokee-Isaac 28B: A 10M-Token Context Agentic Model Built to Run Inside the Customer Boundary
Long-horizon agents accumulate context faster than they resolve tasks. Every tool output, observation, and intermediate reasoning step stays in the window, and the two capabilities that matter — holding that context and staying coherent across it — have so far been available almost exclusively from cloud endpoints. That excludes regulated industries, public-sector institutions, and on-device applications, where the data is not permitted to leave the boundary at all. Pokee AI released Pokee-Isaa
Mistral AI Releases Shieldstral 1.0 3B: An Open-Weights Policy-Adaptive Multimodal Safety Classifier Matching Models 7× Its Size
Mistral AI has released Shieldstral 1.0 3B, an open-weights, policy-adaptive multimodal safety classifier that treats content moderation as a single yes/no question rather than a fixed taxonomy of harm categories. Most guardrail models bake their category list into the weights, so re-targeting one to a new deployment context means retraining — and the same content can be acceptable on a cybersecurity research tool while being harmful on a mental-health platform. Shieldstral inverts that: operat
