How To Use ChatGPT Pro Mode - The Worlds Smartest AI Model

So in this video, I'm going to show you guys exactly what OpenAI just introduced with their OpenAI Pro mode. So currently you can see right here that there are two main offerings from OpenAI. You can see we have the Pro mode right here and we have the plus mode. Now, most people are currently confused with as to why OpenAI are charging $200 a month and they're thinking that this is going to be the new baseline in terms of access to AI. That isn't true. OpenAI didn't communicate this effectively, I don't think, which is why there is so much confusion. So, in the plus tier, you basically get everything. In the three tier, you get extended limits on messaging, file uploads, advanced data analysis and image generation. You also get, you know, the standard and advanced voice mode. But the thing about this mode right here, the $20 per month tier, is that you have limited access. So, for example, you can do everything that you already used to do with GPT4O. The only difference is, is that this isn't actually unlimited. So, for example, if you are messaging 01 and 01 mini, you may actually reach some rate caps. What this means is that if you're having an extended conversation with O1 or O1 mini, what will happen is that over a long period of time, if you keep talking to the model, eventually they might say, okay, your message cap has been reached. Come back in a few days or a few hours. This is something that used to happen due to the fact that this company, OpenAI, doesn't have unlimited compute to be able to give to users paying only $20 a month. So this is something that happens if you're trying to get the most out of 01 and 01 mini. In addition, this is also the same for standard and advanced voice mode. So for advanced voice mode, that's the mode that is really effective at simulating human interaction, especially with a realistic voice. And essentially with this standard advanced voice mode, you only actually have around 45 minutes to an hour per day. So. So if you're someone that loves to use advanced voice mode, understand that there is a time limit in terms of the compute per conversation. So for those of you who are wondering why sometimes it shuts off randomly, that might actually be the case as you have a specific rate limit that you cannot reach. So that is where that price range comes in. Now, of course, we do have $200 a month, and $200 a month is actually quite expensive. Now, I know this is quite expensive because of course it's 10 times the price of $20 a month. But what is actually the difference in terms of the Pro mode versus the O1 and O1 mini? So if we're not going to talk about the Pro Mode, and just based on the tier, you actually get unlimited access to everything. So this means unlimited access to advanced Voice mode, which is basically, of course, as I discussed before, something where you can talk for hours on end with your AI model. You also get unlimited access to 01, 01 Mini and GPT4O. So if you're someone that uses these models for quite some time in terms of writing, and you've got a big project coming up, this is something that allows you to do that. And of course you do get access to O1 Pro mode, which actually uses more compute for the best answers to the hardest questions. And I'm going to explain this in a little bit more detail in a moment, but basically what you have is a system that allows you to get the best answers, the most reliable answers that is currently available today in any AI system. Now, most people will. You won't need access to Pro Mode, especially if you aren't pushing the frontiers in terms of whatever the work it is that you're doing. But it is something that does exist for those who want complete unlimited access to these models. And I'm not surprised they've added this, as I know that there are many users out there who really want unlimited uses of their models, regardless of how much it costs. So now when we actually look at what Pro Mode is, you can see that compared to both O1 and O1 preview, O1 Pro mode performs better on challenging machine learning benchmarks across math, science and coding. So if you're someone that isn't frequently engaging in math, science and or coding, you don't really want to use 01 Pro mode. That means you're not going to be getting the most out of your money if you actually upgrade and you're not engaging in any of these areas. If you're someone that isn't really using these areas, it doesn't really make sense for you to spend $200 a month because that is where this model does excel. You can see that from 01 preview all the way to 01 Pro mode, there is an increased level of ability in terms of Math code and PhD level science questions. Now, some people might argue the fact that when we actually take a look at the jump from 74 to 79, you can see that this jump isn't that different to warrant, of course, a $200 price increase. But that isn't the main thing that OpenAI is actually discussing here. So let me show you guys what that means. Basically what they talk about here is they talk about the fact that this actually makes the model a lot more reliable. So if you're someone who's constantly engaged in those things like science, code and math, you're going to actually make sure you want to get 01 Pro mode, because this is something that is a ton more reliable. They say that to highlight the main strength of the O1 Pro mode, which is improved reliability, they use a stricter evaluation setting. The model is only considered to solve a question if it gets the answer right in four out of four attempts. Four out of four reliability, not just one. And we can see that for four out of four attempts in the competition math, you can see that 01 Pro mode achieves a score of 80. In competition code it achieves a score of 75. And in PhD science level questions it achieves a score of 74. So you can see overall that jump is pretty significant. And you have to remember that because this model is thinking quite a lot. It uses a ton of tokens and all of those tokens do come at the cost of compute, which currently right now is in pretty much a shortage. So if you're someone that has applications to where you're going to be needing your code or needing some math questions, or needing PhD science level questions, this is going to be the model for you if you want that added level of reliability. Because of course, as you know, generative AI is plagued with its ability to not be reliable. Now something else that I do think is pretty interesting is that if you do have image analysis that you do want advanced reasoning on, then O might be something that you want to use. So right now o1 actually has advanced image analysis. So if you actually want to be able to reason with these images, it in an unlimited way, that could be a potential application for you to get 01 Pro mode. I know it is pretty expensive, but this actually allows you the frontier of intelligence and it does think for quite a long time. Now of course, if you want to use 01 Pro mode, I'm going to do a short demo to show you guys exactly what it is. And basically 01 Pro mode is quite different because the answers will take a lot longer to generate and basically there's going to be a progress bar that will send you an in App notification and if you switch away to another conversation. So I'm going to show you guys exactly what you can use 01 Pro mode for and exactly how long it takes to answer certain questions. So firstly, don't laugh at me for spending $200 a month just to ask it this question. But when you ask it any kind of question, it's going to come up with this right here. You can see that it is now thinking about the response and of course you can check the details, but it doesn't have any details available. It should usually have some small reasoning steps right there. But. But this is basically where you'll see a long progress bar come up because the model is going to think really hard about these questions. Now like I said before, 01 Pro mode is not the model that you use when you want to ask a question that is really basic. You can see this question are cats better than dogs? Is something that is super, super simple that I shouldn't have asked this model. But of course I did it to demonstrate exactly how Zero1 Pro mode works. Now if I decide to ask it a much harder question, you're going to see that it takes a lot longer and it's going to be able to provide a longer response. So let's say I wanted to ask it a little bit more of a difficult question that was about reasoning. I can input this prompt from the simple bench benchmark and then we can input this prompt. So you can see here now it is thinking about this. Usually like I said before, it does have some details, but it seems that in 01 Pro mode it doesn't show us any of these change of thought. It just has this area where we can see that the model is thinking. So right now I literally can go over to another tab or to another chat and. And then eventually what we should have is a pop up that pops up and it tells us that okay, look, we have a new chat. The chat was finished, this thing has finished syncing, which is really useful if you're working hard through problems and you don't want to keep of course waiting or want to stay productive. So you can see right here, it pops up and says new chat from OpenAI says finish thinking. And this might not actually be the best demo because in this example it actually didn't output a response, which is rather weird. So one thing that I actually did Forget about the O1 series of models is that you cannot ask these models to think step by step and output your reasoning because. Because OpenAI doesn't want anyone realizing how these models manage to get to their final answer. So if you do ask it a question, you do have to Just ask it that question and just hope that it's able to give you an answer. But of course not output its reasoning step by step. Now you can, if you want to understand why it chose that decision, you can say why would you choose a decision over B decision? But you can't say show your step by step working out. Because I know of course a lot of people who are going to be using these models, are going to be using them on advanced things and they'd love to know why make one decision over a different decision. So you can see right here that it gives us a response, it gives us the pop up and then you can see it manages to determine the reasoning. We also get to see that it reasoned for 1 minute and 15 seconds. It says step by step, determine the number of ice cube placed. Then it outputs this entire thing and it does show us a decent amount of the kind of step by step that it goes through and it gives us the final answer, which is B, which is zero. So overall you can see that this is what you are getting. I know that this isn't largely the best demonstration of how advanced the model is. I don't have any coding examples in math, science or even code for that example. But I'm sure there are many different use cases out there, especially for the advanced users who are really trying to use these models and push them through absolutely everything. So that being said, let me know what you're going to use this model for. I'd love to see your applications and what you're doing with the model. And if you enjoyed the video, I'll see you in the next one.