Sepehr Pakbaz

Yea I get your point but when you have title literally says "o3-mini vs Deepseek-R1" it remind me of those 1000 clickbait videos on youtube that cherry picked few example or prove their point one way or another. I woud say its more helpful if you would have just focused on few areas as opposed to "Some Comparisons to Look Out for : Between o3 Mini and Deepseek R1" or "pricing comparison" or "Choosing Between o3-mini and DeepSeek R1". I see many other articles and again I don't want to be picking on you here but it could have been more coherent and for deepseek I assumed you are referring to the official hosted model which is cloud based which is why I mentioned cloud machine performance comparison at first place.

I am also in favor of open models but when you have an article like this you should at least establish a baseline performance comparison. Even comparing how long it takes on each model on ideal conditions is not correct as we have not established baseline hardware performance. I am sure OpenAI has a significant advantage on the hardware side. For the inference maybe you should have used the hosted version on AWS or Azure to make a more meaningful comparison

You can’t compare models without discussing the performance. DeepSeek R1 usually takes somewhere between 3 to 5 minutes to answer above average questions whereas o3 mini usually takes less than 20 seconds. It’s apple to orange’s comparison