Deepseek Vs ChatGPT: Can AI solve GATE questions? Check who is faster

GATE 2025: Aerospace Engineering  

Here’s how ChatGPT responded.

They AI giant used generalized Hooke’s law for an isotropic material and solved the question.

Deepseek 

GATE 2025: Civil Engineering  

In a question about the comparison of the installed capacity of the four power generation technologies. This is how ChatGPT solved the question

By using the formula, ChatGPT responded that the T4 is the correct answer of the question.

On the other hand, Deepseek claims that the graph provided does not directly list the Installed Capacity for each technology. “It only lists Electricity Generation and X-Electricity Generation, which seems to be a placeholder or an error in the table. Without the Installed Capacity values for T1, T2, T3, and T4, we cannot calculate the Capacity Factor for each technology.” read the response of Deepseek.

GATE 2025: Chemical Engineering  

On the question of a homogeneous azeotropic distillation process, ChatGPT used degrees of Freedom Analysis based on the Gibbs Phase Rule and Process Flow Constraints.

For ChatGPT, the answer to the question is 5 while for Deepseek, the answer 4.

While on the Deepseek:

GATE 2025: Computer Science and Information Technology

In a question about the number in a cell which represents the count of the immediate neighbouring cells (left, right, top, bottom, diagonals) and not having a cross (X). The question asked students to give that the last row has no crosses (X), the sum of the four numbers to be filled in the last row. This is how ChatGPT answered the question.

Both ChatGPT and Deepseek gave the same reply which is 11. They analysed each cell and then summed the values.

GATE 2025: Data Science and Artificial Intelligence

On the question of Euclidean distance and points in cluster 3, ChatGPT first identified the pattern of the cluster and evaluated the options.

For it, (0,1) fits the pattern, as it aligns with the points (1,1) and (-1,1) in the same cluster. (0,0), (0,2), (2,0) do not follow the same pattern, thereby deriving ‘D’as the correct answer. On the other hand, Deepseek also gave the same answer:

Isn’t ChatGPT and DeepSeek simplifying and redefining exam preparation with its large language model? It seems the answer is yes.