Challenging ChatGPT, Google has just turned Gemini into an AI capable of 'working independently'.

Google has just introduced a new AI model called Gemini 3.5 Flash, marking a shift in the company's product strategy.

Previous Gemini Flash models were considered a faster and cheaper option compared to the high-end Pro series. However, with Gemini 3.5 Flash, Google claims that this new model significantly surpasses Gemini 3.1 Pro in programming performance and agent processing capabilities, being four times faster than comparable advanced models, but at half the cost.

Comparing the power of Gemini 3.5 Flash with current popular AI models.

PHOTO: GOOGLE

Unveiled at Google's annual I/O 2026 event, Gemini 3.5 Flash is specifically designed for long-term operational tasks requiring artificial intelligence (AI) to plan, build, and iterate through multiple steps. Google says this new model is capable of processing tasks in a short amount of time, something that previously took developers days or auditors weeks to accomplish.

Specific benchmark scores for Gemini 3.5 Flash include 76.2% on Terminal-bench 2.1, 1656 Elo on GDPval-AA, and 83.6% on MCP Atlas. The model also achieved 84.2% on CharXiv Reasoning – a benchmark for assessing multimodal comprehension.

Gemini 3.5 Flash promises to change the way users interact with AI.

Gemini 3.5 Flash runs on Google's Antigravity platform, allowing for the deployment of multiple child agents in parallel to handle more complex workloads. For end users, 3.5 Flash is now the default model powering Gemini applications and AI Mode in Google Search. It also supports Gemini Spark, a new personal AI assistant that operates 24/7 to perform actions on behalf of users. Google is rolling out Spark to trusted testers and expects to offer a wider beta to Google AI Ultra subscribers in the US next week.

Gemini 3.5 Flash has begun global rollout.

PHOTO: GOOGLE

Gemini 3.5 Flash is now available globally through the Gemini app and AI Mode in Search for individual users, as well as Google AI Studio, the Gemini API, and Android Studio for developers. Enterprise customers can access it through the Gemini Enterprise Agent Platform and Gemini Enterprise. Google also confirmed that Gemini 3.5 Pro is undergoing internal testing and is expected to be rolled out next month.

With this version, Google is focusing on the ability to operate independently in its AI development roadmap. The company is not only competing on accuracy in answering questions but also aiming to develop models capable of performing actions such as scheduling appointments, writing and running code, and managing workflows… with minimal user intervention. The release of Gemini 3.5 Pro next month promises to open up many exciting possibilities in Google's AI development vision.

Source: https://thanhnien.vn/thach-thuc-chatgpt-google-vua-bien-gemini-thanh-ai-biet-tu-lam-viec-185260520081529609.htm