# Modal: SOTA in long-term multimodal AI memory
**URL:**
**Author:**  [@JordanNader](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2FJordanNader)%5C
**Post date:** [November 20, 2025, 4:07pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fmodal-sota-in-long-term-multimodal-ai-memory%2F170676%2F1 "2025-11-20T16:07:11Z")
We’ve been working on a new multimodal memory system called **Modal** , designed to serve as a personalization layer for AI models. It handles multimodal ingestion + retrieval (text, images, audio, video), can be queried in real time, and currently achieves SOTA on the LoCoMo personalization benchmark, outperforming long-context ICL even at 29k tokens.
Early access API is open: [https://www.elicitlabs.ai/early-access](https://www.elicitlabs.ai/early-access)
Benchmark details: [https://elicitlabs.ai/blog/memory](https://elicitlabs.ai/blog/memory)
---
**Author:**  [@system](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fsystem)%5C
**Post date:** [November 20, 2025, 4:07pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fmodal-sota-in-long-term-multimodal-ai-memory%2F170676%2F2 "2025-11-20T16:07:12Z")
Unlisted on November 20, 2025, 4:07pm UTC
---
**Author:**  [@system](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fsystem)%5C
**Post date:** [November 21, 2025, 12:25am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fmodal-sota-in-long-term-multimodal-ai-memory%2F170676%2F3 "2025-11-21T00:25:22Z")
Listed on November 21, 2025, 12:25am UTC