[DigitalToday reporter Yoonseo Lee] Google Gemini 3.1 Pro is a major upgrade from the existing 3.0, delivering performance improvements that amount to a generational shift. As AI is improved to carry out more specific tasks beyond simple answers, prompt design tailored to the objective has become important to fully draw out its potential.

On Feb. 26, local time, IT outlet TechRadar introduced 5 prompts that show Gemini 3.1 Pro's strengths.

First, Gemini 3.1 Pro showed strength in "reverse-checking" that identifies weaknesses in a plan in advance. For example, a user can propose a date plan such as "stargazing on a rooftop after a late dinner," ask it to name 3 points where it could realistically fail, and then rework the schedule with a Plan B for each scenario. Rather than offering a plausible answer, it flags variables often missed during execution and suggests fixes.

In its response, it pointed out that having a late dinner before stargazing could ruin a romantic mood due to weather issues. The schedule was revised into a multi-step plan that included decision points.

This kind of reverse approach is not limited to date plans. It can also be applied to situations where hidden weaknesses can surface later, such as business proposals, travel itineraries or major career changes. The explanation said simulating failure scenarios in advance can help turn vague anxiety into a concrete contingency plan.

Second, Gemini 3.1 Pro can analyse aesthetic elements in video and match them to real locations. For example, it can recommend places in New York state with a similar atmosphere based on Studio Ghibli-style scenery footage. It was assessed as an approach that translates an "animation sensibility" into geographic information, and can be used for various purposes, including finding filming locations and planning theme parties.

Third, Gemini 3.1 Pro supports problem-solving by analysing spaces in images. For example, if a user provides a photo of a half-finished Lego car along with remaining parts, it can suggest assembly methods for a stable structure. There is also room to apply this function to everyday tasks such as furniture assembly or simple home repairs.

Fourth, an SVG rendering function was newly added. Gemini 3.1 Pro can go beyond generating code and implement interactive animations. For example, it can generate an animation of a starry night sky with a rocket flying by. This makes it easier to create customised digital cards or educational content, and allows relatively sophisticated graphic outputs even without extensive development knowledge.

Fifth, Deep Research mode combines hobbies and expertise to provide practical guides. For example, it can research how to convert an existing paperback into a premium hardcover, compiling best practices and work steps. The assessment said that by generating guides that integrate expertise and DIY ideas, AI can be a useful tool for users who want to organise fragmented materials at once.

In this way, Gemini 3.1 Pro is expanding its uses across plan validation, video and image analysis, code and graphics generation, and in-depth research. Accuracy and reproducibility can vary depending on prompt design and the verification process, so users need to make questions specific to their purpose and cross-check the output.