# Meta's Brain2Qwerty v2 Decodes Typed Sentences From Non-Invasive Brain Signals at 61% Word Accuracy

The pipeline, based on magnetoencephalography (MEG), cuts the average word error rate to 39% and reaches 22% for the best participant, but requires a room-scale scanner and per-subject training rather than real-time use.

- Published: 2026-07-18T05:46:36.130Z
- Canonical: https://polylog.news/ai/2026-07-18/meta-s-brain2qwerty-v2-decodes-typed-sentences-from-non-inva
- Publisher: Polylog (AI desk)
- Section: tech
- Sources: [Meta AI](https://ai.meta.com/blog/brain2qwerty-brain-ai-human-communication/), [MarkTechPost](https://www.marktechpost.com/2026/06/30/meta-ai-releases-brain2qwerty-v2-a-non-invasive-meg-brain-to-text-pipeline-decoding-typed-sentences-at-61-word-accuracy/)

Meta detailed Brain2Qwerty, a deep-learning system that reconstructs sentences from MEG recorded while participants type memorized sentences on a QWERTY keyboard. The latest version reaches an average word accuracy of 61%, with an average w…

This story is for subscribers. Read it in full at https://polylog.news/ai/2026-07-18/meta-s-brain2qwerty-v2-decodes-typed-sentences-from-non-inva (subscription information: https://polylog.news/pricing).