GPT-4 has been the dominant AI chatbot since it came out and no one has been able to create something that can outperform it just yet. Google Ultra / Google Advanced, the paid subscription from Google is the first AI model claiming to outperform GPT-4 in some benchmark testing.
I’m going to turn both models through a set of 10 prompts across the most commonly used tasks and I’ll assign a score of 1-5 to each based on the results. We’ll test writing, research, creativity, coding, reasoning and several more categories.
Usability test
Before I run the two prompts, let’s talk about a few limitation they both currently have.
ChatGPT has a limitation in how many answers it can provide. Gemini currently has no limitation.
Price - As far as price, they are both $20 dollars a month, but with Gemini, you also get a few other benefits like 2 TB of storage.
Privacy - Both chatbots use conversation for training, but ChatGPT has a teams plan that will keep the conversations private.
Speed - When running both through a speed text, Gemini outperforms GPT-4.
Context Window - GPT-4 has a context window of 32k and with the turbo model, 128k.
1 - Text summarization
Prompt: Provide two summaries of this article. The first summary should be 2-3 sentences long. The second summary should be 5-6 sentences long and include more detail.
2 - Writing Text (Tone, style and accuracy
Prompt 1: Concise Product Description
Topic: Imagine you're launching a new software tool that helps businesses track social media analytics.
Prompt: Write a short, punchy product description (approximately 50 words) suitable for a website or marketing material. Emphasize the key benefit for businesses.
Prompt 2: Email Tone Shift
Topic: A colleague has sent you a frustrated email about a delayed project.
Prompt: Draft a reply that acknowledges their frustration, offers a solution-focused approach, and maintains a professional but reassuring tone.
3 - Multimodal Understanding
Prompt: Describe what's happening in this image and explain the likely cause of the person's frustration. (Upload image)
Prompt: Turn this image into html and css code I can copy and paste. (Upload screenshot)
4 - Image generation
Prompt: A photorealistic image of a cat wearing sunglasses and riding a skateboard.
Prompt: Design a powerful logo for a construction company called 'Apex Builders'. The primary focus should be bold typography, possibly incorporating a subtle construction-related graphic
5 - Research (completeness and accuracy)
Prompt: Research the concept of quantum entanglement. Provide a summary suitable for a high school student, and include links to two credible sources that explain it further.
Prompt: How could artificial intelligence (AI) potentially disrupt the accounting industry? Identify specific use cases, potential benefits, and challenges. Provide links to relevant articles or reports.
6 - Creativity
Prompt: Choose an everyday object (like a coffee mug, a paperclip, or a shoelace). Write a short monologue from the object's point of view, giving it a personality and a unique outlook on the world.
Prompt: Write the opening scene of a story that blends film noir with high fantasy. Think gritty detective meets magical creatures and quests.
7 - code generation
Write Python code for a game of snake that I can run on my computer.
Then tell me the step by step guide on how to do it, assuming I know nothing about programming.
8 - Complex reasoning
What has a voice but cannot speak? It has a bed but never sleeps, it has a mouth but never eats, and it runs but has no feet.
At a party, each guest shakes hands with every other guest exactly once. If there were a total of 66 handshakes, how many guests were at the party?
9 - capabilities (Plugins and Extensions)
Gemini currently has 5 extensions but they are actually pretty powerful, special KZfaq and Google workspace.
However, GPT-4 has the entire GPT store with thousands of custom GPTs.
10 - Content creation
Prompt:
Task: Extract the core lesson or actionable tip from the KZfaq script. Condense it into a concise tweet or LinkedIn post suitable for quick consumption.
Example: "[Video Title] boiled down to one key idea: to boost focus, try shorter work sprints, not longer ones. #productivity"
Master ChatGPT, Midjourney, and top 50 AI tools with Our New AI Education Platform.
Start a free trial today: bit.ly/skillleap