Artificial Intelligence

Claude 3.5 vs. ChatGPT4o: Which AI assistant wins?

Choosing the suitable AI assistant comes down to specific needs, to help you decide, we’ve compared both models side by side.

image credit: bendada.com, ChatGPT4o vs Claude 3.5

As the competition in the generative AI market increases, more AI models are being released daily, but only a few excel in roundedness. So far, two models have stood out in their ability to understand instructions and carry out complex tasks: OpenAI’s ChatGPT and Anthropic's Claude.

ChatGPT, launched in 2022, is popular and has attracted millions of consumers and businesses for its powerful language abilities, extensive knowledge base, and user-friendly interface. Claude, Anthropic’s ChatGPT rival recently joined the AI race. Users can explore different versions, such as Claude’s Sonnet 3.5 and ChatGPT’s GPT-4omni, arguably the most powerful LLMs in 2024.

In this case, ‘best’ simply refers to ‘best fit for me.’ So to help you decide, we’ve compared both models in terms of capabilities, strengths, weaknesses, technical details, privacy, and overall features.

Let’s get started!

Different Claudes

Claude is primarily a text-based LLM, though it can generate charts and infographics. In May 2024, Anthropic released a major update to Claude 2. The following month, they launched Claude 3.5, an enhancement of the mid-tier Sonnet model, Claude 3 Opus. Anthropics Benchmark tests indicate that Claude 3.5 outperforms the Opus and many competitor models.

Claude is available as an app for Android and iOS and its web platform. Alongside a limited free edition, Anthropic offers two paid subscription tiers: ₦32,000/$20 for individuals and ₦48,000/$30 for teams of at least five subscribers.

Claude 3 models are accessible via Amazon Bedrock and Google Cloud Vertex AI-managed service platforms.

ChatGPT

OpenAI, on the other hand, offers a broader range of models than Anthropic, including various versions of GPT in its larger model lineup. For instance, GPT-4o is a multimodal model that processes text, images, and audio. ChatGPT is accessible online at https://www.chatgpt.com, as an Android or iOS app, and through OpenAI’s developer API.

Claude 3.5 vs. GPT4o: Features compared

While ChatGPT excels in analytical tasks, Claude now has a limited data analysis feature. In content generation, both have areas where they excel and based on your preference, either language model could do the job for you.

Here’s a quick table summarising features between ChatGPT and Claude

LLM	Claude	ChatGPT
Creator	Anthropic	OpenAI
Language Models	Claude Haiku, Claude 3.5, Clause Opus	GPT-3.5, GPT-4, GPT-4o
Languages	Over 12 languages including English, French and German	90+ languages
Image Generation	No	Yes
Image-based responses	Yes	Yes
Voice Chat and Text to speech	No	Yes
Tone and Engagement	Fluid, life-like, articulate	Requires specific prompts to get the desired tone of voice
Logic	Excels in ethical reasoning, and can also provide logical explanations	Sound Logic, might not always provide detailed or tailored answers
Data Processing	Able to analyse long documents and summarise key points and insights	Could struggle with large amounts of bulk data, prompts are broken down and entered one by one to generate better results
Data Visualisation	Centred around VizGPT, it allows you to build complex visualisations step by step using a chat interface	It can directly generate charts and graphs through its code interpreter
Context window	Its larger context windows allow for coherence over extended conversations, which allows for a better understanding of long texts( around 200,000 tokens)	Good context window but may lose track of context over long exchanges (around 32,000 tokens)
Memory	No persistent memory, each transaction is self-contained	Persistent memory helps the language model remember details from chats, providing more personalised results over time
Customisation	It operates based on its training, with no options for customisation	Custom instructions allow users to personalise responses by providing preferred response styles, and details enabling the users to create custom GPTs
Code Generation	Better at generating optimised code for data science tasks, and can explain the code with insights.	Good at code generation but may only sometimes provide insights or analysis.
Pricing	Claude 2 is free with 3A going for $20/month which gives you access to Claude 3.5	ChatGPT Plus costs $20/month which gives you access to chatGPT 4

Some other features of these LLMs that some users could find useful include:

ChatGPT’s browsing feature allows your language model to surf the web in a certain direction and make more streamlined predictions. This is a beta feature in GPT4o which Claude does not have.

GPT-4o Beta Features

Aside from being better for step-by-step tasks, Claude 3.5 supports multiple chats in a single interface, which is a feature ChatGPT needs to include. Claude excels in open-ended tasks where the project's final direction could be clearer.

While Claude is efficient and performs tasks remarkably well, users often complain about its chat limit. The free version restricts the number of chats, triggering a pause before functions resume. This limitation can be particularly inconvenient when running code, as there's no way to return to the last progress if the chat limit is reached.

Chat limits on Claude

While ChatGPT also has a chat limit, free chats have a higher cap than Claude.

Claude 3.5 vs. ChatGPT4o: How to choose the right AI assistant

Large Language Models (LLMs) like Claude 3.5 and ChatGPT 4.0 are trained on massive datasets to perform various tasks. When choosing between the two, consider your specific needs.

Say you're brainstorming marketing ideas, and you need to go through large volumes of data to make decisions, or predictions, would you use Claude 3.5 or ChatGPT4o?

Data Analysis:

Because Claude 3.5 features a large context window, it's perfect for handling extensive text data. You can upload lengthy documents or market research reports, and your AI assistant will process and learn from them, making streamlined predictions easy. You're better off using the Claude if you have a large array of data to sift through.

Claude analyses and summarises large texts

Image Generation and Manipulation:

Claude 3.5 is a text-based AI chatbot, it does not have capabilities to generate or manipulate images like chatPT4o does, so if you’re looking for an AI assistant that could develop, edit or manipulate images or videos, GPT4o is your guy.

GPT4o generates images from prompts

Code Generation:

While users argue that both models can generate code, Claude 3.5 generally outperforms ChatGPT 4.0 in terms of speed and code quality. It also offers more granular control over code complexity.

Claude 3.5 generating code based on prompts

Claude 3.5 generates code with a tweakable feature and lets you customise your code, GPT-4o doesn’t.

Organisation and Search:

Some users consider the chat archive search feature in Claude to be a flaw. For instance, it doesn't allow you to preview a project when you hover over it, displaying only the name and description. This makes it difficult to find specific conversations for reference. As you accumulate projects and chats, easily looking back and referring to them becomes essential.

Bottom Line: ChatGPT 4.0 currently provides a more user-friendly chat archive, allowing users to find and reference past conversations. Claude 3.5's search functionality could be improved.

Claude 3.5 vs. ChatGPT4o: Privacy and Security

Claude does not automatically retain user data like ChatGPT does. Using Claude, your chat is only used for training if you flag it with one of those thumbs-up/down buttons.

Your conversations are flagged for a “Trust & Safety review, " meaning they indirectly use the signal from that review process for "model training.

Claude 3.5 vs. ChatGPT4o: Updates

Claude and ChatGPT, have undergone significant evolution since they launched. Claude has advanced to the Claude 3 family. Anthropic has enhanced Claude.ai for Pro and Team users by introducing project functionality, enabling better chat organisation, and integrating organisational knowledge.

At the same time, ChatGPT has progressed through iterations like GPT-3.5 and GPT-4 there is tremendous potential for creating even more advanced language models, especially in context awareness. Another area of development for ChatGPT is related to multi-modal learning

These systems represent major strides in conversational AI, with ongoing improvements in language understanding, task performance, and knowledge expansion.

Our verdict

In reasoning tests, Claude 3.5 and GPT-4o tied, particularly in logic. However, GPT-4o leads in mathematical reasoning, an area where Claude could improve. Claude excels in generation speed, producing text and code faster than GPT-4o. ChatGPT-4o stands out for its integrations and flexibility, making it a more versatile tool.

Claude-3.5-Sonnet offers quicker responses and detailed explanations, while GPT-4o shines in algorithmic tasks, image generation/manipulation, and performance optimisation. Additionally, ChatGPT allows for external API calls, enabling features like saving conversation results to a database or triggering other apps.

Claude 3.5 and ChatGPT-4o perform similarly based on hands-on testing, with each tool having unique advantages depending on the task and user preferences.

FAQs: Claude 3.5 vs. ChatGPT4o

Which AI assistant is better: Claude 3.5 or ChatGPT4o?

Both are excellent AI assistants. Claude 3.5 excels in creative projects like writing and editing, has better coding abilities, and has cheaper API costs. For image generation and manipulation, GPT4o is the better choice.

Which AI assistant is better for generating code?

Claude 3.5, with its wider context window and superior command understanding, is better for creating large amounts of code by reference than GPT4o.

Can Claude 3.5 interpret images?

Yes, Claude 3.5 can interpret image prompts effectively.

Can Claude 3.5 generate images?

Unfortunately, Claude 3.5 cannot generate images. GPT4o is better for image generation and manipulation.

Can I use Claude or GPT4o for free?

Both Claude 3.5 and GPT4o are free, with varying message limits based on demand. Claude notifies you when you reach your daily limit, which resets every morning. Claude's free chat session limit is shorter than ChatGPT's.