88Z AI, 13 November — Looki has released its self-developed multimodal model series, Loom, focusing on first-person life understanding and long-term memory. The series includes Loom Sense, Loom Moment, and Loom Memory, designed to optimize continuous perception, event segmentation, and cross-time memory retrieval. It significantly reduces inference costs and improves processing efficiency, turning fragmented observations into traceable personal context.
At a glance
- The Loom series connects perception to memory pipelines, reducing inference costs by approximately 94%
- Loom Moment outperforms open-source models in segmentation, while Memory achieves a comprehensive score of 79

The update
Looki has officially launched Loom, a self-developed multimodal model series centered on first-person life understanding and long-term memory construction. This model connects the entire pipeline from continuous perception and experience organization to cross-time memory retrieval, aiming to weave fragmented observational data into understandable and traceable personal context. In terms of performance, Loom Sense achieves an F1 score of 88.6 for event judgment, with a single-window processing time of only 0.51 seconds and an approximate 94% reduction in inference costs. Loom Moment outperforms open-source models in segmentation tasks, offering more accurate evidence attribution. Loom Memory scores 79 comprehensively, with information processing volumes at 14.5% and 30.5%, demonstrating efficient information handling capabilities.
Sources & editorial note
This article is an AI-assisted synthesis by 88Z AI based on the public sources below. Product claims and availability are attributed to their publishers. It does not represent independent testing or first-hand reporting.
- LookiAISource published
Source dates above refer to the original announcements; the date under the headline is this article’s publication on 88Z AI.