Google just launched its updated version of the Gemini 2.0 family of AI models.
Called Gemini 2.0 Flash, this is up for grabs in an exciting chat version that users around the world can experiment with. Meanwhile, developers were also treated with another trial of its multimodal version. This features both text-to-speech conversion and features for image generation.
As per CEO Sundar Pichai, Gemini 1.0 was more related to keeping data organized and comprehending information. This latest model on the other hand features more useful offerings. It’s also seen outperforming previous models in terms of user request and code generation. It can also provide more factually correct replies. However, one weak point where it’s subpar in performance to Gemini 1.5 Pro is comprehending longer contexts.
To attain access to chat-optimized variants of experimental Gemini Flash 2.0, users get the chance to choose it by seeing the model’s drop-down menu available on the mobile web as well as the desktop. It’s going to be up for grabs on the Gemini mobile app soon, the firm adds.
The latest multimodal variant of Gemini Flash 2.0 is going to be found on Google’s AI Studio as well as Vertex AI developer apps. Overall availability of Gemini 2.0 Flash’s multimodal variant will appear in January. This is where users will find different Gemini 2.0 model sizes. More plans for expansion to other Google domains are on the horizon by early next year.
This is clearly Google’s main effort in the tech world’s highly competitive AI race. There are plenty of archrivals including Microsoft, OpenAI, and Meta who have similarly robust offerings.
Alongside this release, Google did tease about other models that are solely designed to assist with the likes of research and hence are more agentic AI systems. They comprehend more of the real world, plan several steps ahead, and even take action on the user’s behalf. Of course, supervision is necessary.
Last week we saw Google’s CEO speak about challenging Microsoft’s latest AI advancements. This includes a side-by-side competition to determine who has the best model. This just goes to show how confident he is about winning over the software giant.
Image: DIW-Aigen
Read next: New Report Suggests YouTube is Pushing Weight Loss and Eating Disorder Content to Young Girls
Called Gemini 2.0 Flash, this is up for grabs in an exciting chat version that users around the world can experiment with. Meanwhile, developers were also treated with another trial of its multimodal version. This features both text-to-speech conversion and features for image generation.
As per CEO Sundar Pichai, Gemini 1.0 was more related to keeping data organized and comprehending information. This latest model on the other hand features more useful offerings. It’s also seen outperforming previous models in terms of user request and code generation. It can also provide more factually correct replies. However, one weak point where it’s subpar in performance to Gemini 1.5 Pro is comprehending longer contexts.
To attain access to chat-optimized variants of experimental Gemini Flash 2.0, users get the chance to choose it by seeing the model’s drop-down menu available on the mobile web as well as the desktop. It’s going to be up for grabs on the Gemini mobile app soon, the firm adds.
The latest multimodal variant of Gemini Flash 2.0 is going to be found on Google’s AI Studio as well as Vertex AI developer apps. Overall availability of Gemini 2.0 Flash’s multimodal variant will appear in January. This is where users will find different Gemini 2.0 model sizes. More plans for expansion to other Google domains are on the horizon by early next year.
This is clearly Google’s main effort in the tech world’s highly competitive AI race. There are plenty of archrivals including Microsoft, OpenAI, and Meta who have similarly robust offerings.
Alongside this release, Google did tease about other models that are solely designed to assist with the likes of research and hence are more agentic AI systems. They comprehend more of the real world, plan several steps ahead, and even take action on the user’s behalf. Of course, supervision is necessary.
Last week we saw Google’s CEO speak about challenging Microsoft’s latest AI advancements. This includes a side-by-side competition to determine who has the best model. This just goes to show how confident he is about winning over the software giant.
However, not all things are as smooth as they appears from the above, as Google's Gemini for Google Workspace, integrated into tools like Gmail and Docs, faces challenges in competing with rivals such as Microsoft Copilot and ChatGPT. Despite offering features like email rewriting and document summarization, internal data shows Gemini lags in brand recognition and usage, particularly in the US and UK. However, the model performs better in markets like Brazil and India. Google has seen improvements since transitioning from "Duet AI" to "Gemini" and remains optimistic about growth in 2025, despite leadership changes within the Workspace team.
Read next: New Report Suggests YouTube is Pushing Weight Loss and Eating Disorder Content to Young Girls
