TL;DR
- Before Dimensity 9400 SoC, multimodal Gemini Nano was only available on the Pixel 9 series.
- Dimensity 9400 is the first chip to support on-device multimodality after Google’s in-house chip.
- The “Dimensity 9400 features a new 8th generation NPU with hardware acceleration for text, image, and speech.”
The newly-launched Dimensity 9400 SoC comes with a lot of improvements in its processing and graphical prowess. However, the most exciting development seems to be the inclusion of on-device, multimodal Gemini Nano. This marks a big step in the Android smartphone industry, as the Dimensity 9400 SoC is the only processor (besides Tensor G4) to achieve this.
ALSO READ: Vivo X200 Pro Vs. X200 Pro mini: Official Images Reveal Differences In Design And Colors
What Is Google’s Gemini Nano?
Released in December 2023, Gemini Nano is a part of Google’s AI models family: Gemini. Compared to the other models in the lineup, Nano is smaller and lighter, making it the ideal candidate for use on devices with relatively less processing power, like smartphones.
Text-Based Vs. Multimodal AI Model: What’s The Difference?
In its native form, Gemini Nano enables features like text generation, summarization, audio transcripts, etc. However, the latest version of the AI model supports multimodality, implying that it can process inputs in text, images, audio, and videos. Last but not least, Gemini Nano is a multilingual language model: it can understand and create content in multiple languages.
While the model was initially available on Google’s Pixel lineup of smartphones (including the Pixel 8 and Pixel 9 series), several OEMs have adopted Gemini Nano in the last year. Today, the AI model facilitates many smartphones, including the Samsung Galaxy S24 series, the Galaxy Z Fold 6, Galaxy Z Flip 6, Motorola Edge 50 Ultra, Razr 50 Ultra, and the Realme GT 6, among others.
ALSO SEE: AMOLED Display Mobile Phones Under 20000
Dimensity 9400 Supports Multimodal Gemini Nano: Here’s What It Means
As mentioned previously, there are two types of Gemini Nano models. While one understands only text-based inputs, the other can make sense of images, videos, and audio, making it multimodal.
Before Dimensity 9400 SoC, multimodal Gemini Nano was only available on the Pixel 9 series, including the vanilla Pixel 9, Pixel 9 Pro, Pixel 9 Pro XL, and the Pixel 9 Pro Fold. Google optimized the Tensor G4 SoC to run the AI model locally on the devices, which enabled quicker processing.
Dimensity 9400 SoC To Power Multimodal Gemini Nano On Non-Google Flagships
With the advent of the Dimensity 9400 SoC, the multimodal Gemini Nano is no longer exclusive to the Tensor G4 SoC. This is the first chip to support on-device multimodality after Google’s in-house chip. As OEMs drop their Dimensity 9400-powered smartphones, people will be able to use all the advanced features it enables, like Pixel Screenshots, Pixel Recorder, Phone by Google, etc.
The “Dimensity 9400 features a new 8th generation NPU with hardware acceleration for text, image, and speech,” writes MediaTek in an official release. “By enabling multimodal models, users will be able to take images and receive detailed descriptions of what’s been captured,” adds the company.
ALSO SEE: Bluetooth Calling Smartwatches Under 1000
However, whether companies will provide the features as is (which would require them to work closely with Google) or develop custom versions of them is something that we’re most interested in seeing. Most recently, Google has allowed developers to experiment with the text-based Gemini Nano to use the AI model’s capabilities in their apps.
You can follow Smartprix on Twitter, Facebook, Instagram, and Google News. Visit smartprix.com for the latest tech and auto news, reviews, and guides.