# Titans + MIRAS: Helping AI have long-term memory

DevFeed: [Titans + MIRAS: Helping AI have long-term memory](<https://devfeed.tech/articles/titans-miras-helping-ai-have-long-term-memory-6900.md>)

Original publisher: [Read original article](<https://research.google/blog/titans-miras-helping-ai-have-long-term-memory/>)

Published: 2025-12-04T19:26:09Z

Content type: article

Language: en

Sources: [The latest research from Google](<https://devfeed.tech/sources/the-latest-research-from-google.md>)

Topics: [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Generative AI](<https://devfeed.tech/topics/generative-ai.md>), [Transformer architecture](<https://devfeed.tech/topics/transformer-architecture.md>), [Google](<https://devfeed.tech/topics/google.md>), [Mamba](<https://devfeed.tech/topics/mamba.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>), [Compression](<https://devfeed.tech/topics/compression.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [compression](<https://devfeed.tech/tags/compression.md>), [context](<https://devfeed.tech/tags/context.md>), [generative-ai](<https://devfeed.tech/tags/generative-ai.md>), [machine-intelligence](<https://devfeed.tech/tags/machine-intelligence.md>), [mamba](<https://devfeed.tech/tags/mamba.md>), [natural-language-processing](<https://devfeed.tech/tags/natural-language-processing.md>), [neural-networks](<https://devfeed.tech/tags/neural-networks.md>), [offline](<https://devfeed.tech/tags/offline.md>), [real-time](<https://devfeed.tech/tags/real-time.md>), [recurrent](<https://devfeed.tech/tags/recurrent.md>)

## AI overview

Google Research introduces Titans and the MIRAS framework, which combine recurrent-model speed with transformer-like accuracy for handling very long contexts. The approach uses test-time memorization and an adaptive neural long-term memory that updates as data streams in, without offline retraining.

## Source excerpt

Generative AI