Top News

Tezzbuzz Desk- Google has launched a new and lightweight AI model for its AI coding platform. The company introduced the “Gemini 3.5 Flash Low” variant on Tuesday, May 26, which has been specially designed keeping in mind lower token spend and better usage experience.

The company says that the new model will save users from the problem of quickly reaching the rate limit and will help in using AI tools for a long time.

Why did a new model have to be introduced?

Recently, Google removed the message based system on its AI platform and implemented a compute based usage model. After this, many users had complained that tokens are getting exhausted very fast even in small and simple tasks. Considering this problem of the users, the company has prepared a new “Gemini 3.5 Flash Low” model.

New model will spend 45% less tokens

While giving information on social media platform X, Varun Mohan associated with Google DeepMind said that the new Flash Low model will use about 45 percent less tokens than before.

According to the company, despite the low token cost, no major compromise has been made in the performance of the model. Especially in software engineering and everyday AI tasks, it will give better results.

Flash model now divided into three parts

Google has now divided its Gemini 3.5 Flash model into three different variants—

• Gemini 3.5 Flash Low – For light and everyday tasks
• Gemini 3.5 Flash Medium – for general use
• Gemini 3.5 Flash High – For complex and heavy AI tasks

The company says that with this, users will be able to choose the right model according to their needs.

Free users also got benefit

Google not only launched the new model, but also reset the Gemini quota of all plans this week. Free users will also benefit from this. The company wants users to be able to complete their AI projects without interruption and to reduce the problem of quickly exhausting the token limit.

Questions also raised on image generation limit

Some users also complained about the image generation limit. One user claimed that 1000 images can be generated on Codex, whereas only 24 images are being generated in Antigravity’s Ultra plan.

On this, Varun Mohan admitted that the current limit is very low and there is a need to increase it. However, no official announcement has been made yet regarding the new limit.

Google’s big step in the race of AI tools

Amidst the increasing competition in the AI ​​sector, Google is continuously working on improving its Gemini models. The new Flash Low variant can be a relief for those users who want to use AI tools for a long time but are troubled by the token limit.