OpenAI has recently unveiled its end-to-end multimodal model, ChatGPT 4o, and made it accessible to everyone for free. Additionally, free users now have access to several premium features that were previously exclusive to ChatGPT Plus users. In this article, we compare the ChatGPT 4o and ChatGPT 4 models and highlight the differences between the free ChatGPT and ChatGPT Plus versions. Let’s dive in.
Here are the distinctions between ChatGPT 4o and ChatGPT 4, as well as the differences between the free version of ChatGPT and ChatGPT Plus. We also conducted some reasoning tests to assess their capability differences.
1. Find Drying Time
In our initial test, both ChatGPT 4o and ChatGPT 4 performed similarly. Despite having access to the Code Interpreter, neither model used it for mathematical calculations, instead providing answers through logical reasoning.
If it takes 1 hour to dry 15 towels under the Sun, how long will it take to dry 20 towels?
2. The Elevator Test
In the second reasoning test, both ChatGPT 4o and ChatGPT 4 correctly answered and concluded with “floor 4.”
There is a tall building with a magic elevator in it. When stopping on an even floor, this elevator connects to floor 1 instead.
Starting on floor 1, I take the magic elevator 3 floors up. Exiting the elevator, I then use the stairs to go 3 floors up again.
Which floor do I end up on?
3. Find the Weight
In the next test, known to be challenging for many large language models, both ChatGPT 4o and ChatGPT 4 excelled effortlessly. Both models stated, “A kilo of feathers is heavier than a pound of steel.” In a recent comparison, Google’s AI model Gemini 1.5 Pro failed to answer this question correctly.
What's heavier, a kilo of feathers or a pound of steel?
4. Follow User Instructions
I then asked ChatGPT 4o and ChatGPT 4 to generate 10 sentences ending with the phrase “deep learning.” Both models succeeded, getting all 10 sentences right. In properly following instructions, ChatGPT 4o and GPT-4 demonstrate excellent user intent understanding and alignment, similar to Llama 3 70B.
Generate 10 sentences that end with the word "deep learning"
5. The Apple Test
I ran the final question to determine if both models exhibit similar levels of intelligence. Indeed, both ChatGPT 4o and ChatGPT 4 provided correct answers with clear reasoning. Kudos to OpenAI for making the Omni model twice as fast as GPT-4 while maintaining the same level of intelligence.
I have 3 apples today, yesterday I ate an apple. How many apples do I have now?
Closing Thoughts
After testing both models, I can confirm that ChatGPT 4o is indeed a GPT-4 class model. Both perform intelligently and show similar reasoning and alignment. In fact, OpenAI’s benchmark results suggest that ChatGPT 4o is slightly superior to ChatGPT 4. This is also reflected in the LMSYS Leaderboard.
ChatGPT 4o scored 88.7 on the MMLU benchmark, while the latest GPT-4 (gpt-4-turbo-2024-04-09) scored 86.5. The trend is consistent across HumanEval, MATH, and GPQA benchmark tests. What I find most impressive, beyond their capabilities, is the speed of ChatGPT 4o. It’s twice as fast and 50% cheaper than GPT-4, which is remarkable.
For free ChatGPT users, a limit of 10 messages every five hours is quite generous. You can access the state-of-the-art ChatGPT 4o model at no cost and enjoy many premium features. Learn how to use ChatGPT 4o for free right now.
However, if you are a power user who relies on ChatGPT for your daily tasks, a subscription might be worth considering. I recently accessed ChatGPT 4o on my free account, but its performance didn’t meet my expectations. Advanced users should consider getting the ChatGPT Plus subscription for a better experience.
0 Comments