Future

Merge

Neuralink

Meta

<ul>
 <li>Brain2Qwerty v2 decodes sentences from non-invasive brain recordings in real time.</li>
 <li>Meta trained the system on 22,000 sentences from nine volunteers.</li>
 <li>The project includes open code, datasets, and research funding support.</li>
</ul>
Brain2Qwerty v2 marks Meta’s latest breakthrough in non-invasive brain-to-text technology. The new artificial intelligence system converts raw brain signals into written sentences without requiring surgery. Meta says the research could eventually help millions of people living with brain injuries or neurological disorders that prevent normal communication.
<h2 class="wp-block-heading" id="brain2qwerty-v2-improves-accuracy-with-ai-models">Brain2Qwerty v2 Improves Accuracy With AI Models</h2>
Unlike traditional brain-computer interfaces that rely on implanted electrodes, Brain2Qwerty v2 uses a helmet-like magnetoencephalography (MEG) scanner to capture brain activity. Researchers trained the model on approximately 22,000 sentences collected from nine volunteers, each participating in around 10 hours of typing sessions while wearing the device.
<figure class="wp-block-embed is-type-rich is-provider-x wp-block-embed-x">
 <div class="wp-block-embed__wrapper">
 <blockquote class="twitter-tweet">
 We’re sharing the next major milestone in our non-invasive brain-to-text decoder research: Brain2Qwerty v2. Building on v1, which was published today in <a href="https://x.com/Nature?ref_src=twsrc%5Etfw" rel="noopener noreferrer nofollow">@Nature</a>, Brain2Qwerty v2 is the highest-performing end-to-end pipeline capable of real-time sentence decoding from raw brain… <a href="https://t.co/qJ8qrrgTaF" rel="noopener noreferrer nofollow">pic.twitter.com/qJ8qrrgTaF</a>— AI at Meta (@AIatMeta) <a href="https://x.com/AIatMeta/status/2071566924803395741?ref_src=twsrc%5Etfw" rel="noopener noreferrer nofollow">June 29, 2026</a>
 </blockquote>
 </div>
</figure>
The AI system processes raw neural signals through an end-to-end deep learning model. Meta also fine-tuned large language models using neural data, allowing the system to understand semantic context instead of decoding individual characters alone.
According to Meta, Brain2Qwerty v2 achieved an average 61% word accuracy, a significant improvement over earlier non-invasive methods that reached only about 8%.
<h2 class="wp-block-heading" id="brain2qwerty-v2-supports-future-brain-computer-interfaces">Brain2Qwerty v2 Supports Future Brain-Computer Interfaces</h2>
Meta believes Brain2Qwerty v2 narrows the performance gap between non-invasive systems and surgically implanted brain-computer interfaces. The company noted that decoding accuracy continued improving as additional training data became available, suggesting future versions may perform even better.
To encourage further research, Meta is releasing the project’s source code and dataset through its Digital Brain Project. The company also announced a $5 million fund dedicated to expanding open neuroscience datasets.
The research appears alongside a paper published in Nature Neuroscience, where Meta argues that safer, non-invasive technologies could make brain-computer interfaces more practical for wider medical use. Similar efforts are underway across the industry, including Neuralink and Merge Labs, which are developing communication technologies for people affected by neurological conditions.
<pre class="wp-block-preformatted">Disclaimer: This article is for informational purposes only and does not constitute financial advice. CoinCryptoNewz is not responsible for any losses incurred. Readers should do their own research before making financial decisions.</pre>
&lt;p&gt;The post <a href="https://coincryptonewz.com/brain2qwerty-v2-hits-61-word-decoding-accuracy/" rel="noopener noreferrer nofollow">Brain2Qwerty v2 Hits 61% Word Decoding Accuracy&nbsp;</a> first appeared on <a href="https://coincryptonewz.com" rel="noopener noreferrer nofollow">Coin Crypto Newz</a>.&lt;/p&gt;

Brain2Qwerty v2 Hits 61% Word Decoding Accuracy

Table of Contents

CoinCryptoNewz

Other articles published on Jul 1, 2026

Join our free newsletter for daily crypto updates!