As someone who spends every day testing the world’s smartest chatbots, I’ve learned that not all models are great at creative work. While many can write well, most only touch the surface in terms of tone and creativity. Coding and summarizing are far more analytical, so it makes sense that chatbots can think like pros, but when it comes to something that needs heart, tone and imagination, most chatbots stumble.
Claude is known for being one of the most creative chatbots and with ChatGPT-5.1 now available to ChatGPT Plus subscribers, I just had to know how they compare. So I decided to run the two AIs through 9 creative prompts across writing, brainstorming and storytelling to see which one felt like the better creative partner.
Both are fast, intuitive and wildly capable — but only one made me feel like I was co-creating with another mind, not just prompting a machine.
1. Hooking readers
Prompt: “Write the first paragraph of a novel about a woman who time-travels through her old text messages.”
ChatGPT-5.1 presented a clever, high-concept premise where the phone’s physical damage literally cracks time open, turning memories into portals.
Claude 4.5 Sonnet crafted an emotionally resonating scene by establishing immediate mystery with a specific, impossible message from the dead.
Winner: Claude wins for creating a more powerful emotional anchor and a deeper sense of immediacy. What Claude delivered felt more like the true beginning of a novel.
2. Ideas for a podcast
Prompt: “Come up with five new podcast concepts for a journalist who covers AI.”
ChatGPT-5.1 provided creative and well-named ideas that are engaging and thoughtfully framed, with a slightly more mainstream, listener-friendly tone.
Claude 4.5 Sonnet delivered a list of exceptionally strong, commercially viable concepts that feel fresh and directly serve a tech journalist's beat with clear audience appeal.
Winner: ChatGPT wins for concepts that are entertaining and relevant to what listeners most want to know about without going too deep into the AI space, potentially alienating listeners who aren’t familiar with all the current AI tools.
3. Creating a TikTok script
Prompt: “Create a 30-second TikTok script about the future of work and AI.”
ChatGPT-5.1 generated a clear, informative and reassuring script that directly addresses common anxieties with a strong, empowering message.
Claude 4.5 Sonnet created a highly engaging, character-driven script that uses visual comedy and a relatable scenario to make its point about the future of work.
Winner: Claude wins for its superior understanding of the TikTok medium. The script was the most creative and better embodies the show-don't-tell, entertainment-first ethos of a viral TikTok.
4. Developing a plot
Prompt: “Suggest three plot twists for a cozy mystery set in the suburbs.”
ChatGPT-5.1 gave me three clever, high-concept twists that are perfectly suited to a modern suburban setting, blending classic mystery tropes with contemporary elements like AI and identity theft.
Claude 4.5 Sonnet crafted three deeply psychological and character-driven twists that subvert the genre by challenging the very nature of guilt, justice and the detective's role.
Winner: Claude wins for creating twists that fundamentally reshape the entire narrative and its moral compass. While ChatGPT's ideas are fun and fitting, Claude's concepts are more sophisticated and original, proving to be more creative.
5. Hype machine
Prompt: “Write a short pep talk for a parent on a tough morning.”
ChatGPT-5.1 provided a warm, relatable and gently humorous pep talk that feels like a supportive friend offering a quick, comforting boost.
Claude 4.5 Sonnet delivered a deeply empathetic and emotionally resonant pep talk that acknowledges the parent's struggle with raw honesty before building them up with profound validation.
Winner: Claude wins for better depth and psychological insight. While ChatGPT's response is kind and effective, Claude's feels more like a conversation, meeting the parent in their exhaustion and leaving them feeling truly seen and fortified. In other words, a more creative and in-depth assistant.
6. Writing taglines
Prompt: “Create a tagline for a new productivity app that feels playful, not corporate.”
ChatGPT-5.1 made a list of concise, positive and universally relatable taglines that feel like a supportive friend, with "Your to-do list's new BFF" being a particularly strong, playful concept.
Claude 4.5 Sonnet offered a list of taglines that are witty, culturally savvy and gently subversive, directly mocking corporate hustle culture with lines like "Productivity without the existential dread."
Winner: ChatGPT wins for better understanding of brevity and creativity for solid taglines.
7. Explaining humor
Prompt: “Explain what makes something funny — and write an example.”
ChatGPT-5.1 was concise and delivered an explanation centered on the core principle of surprise, supported by a relatable and effective example that clearly illustrates the concepts.
Claude 4.5 Sonnet provided a comprehensive, structured breakdown of comedic theory, complete with six distinct mechanics and a brilliantly crafted original example that demonstrates all of them in action.
Winner: Claude wins for deconstructing the question into a teachable framework, providing a veritable toolkit for understanding and creating comedy.
8. Attention to detail
Prompt: “Describe a futuristic kitchen using sensory detail only (no tech jargon).”
ChatGPT-5.1 painted a minimalist vision of a futuristic kitchen, focusing on ambient light, subtle scents, and seamless automation to create a serene, almost magical atmosphere.
Claude 4.5 Sonnet immersed us into the scene with a deeply sensory narrative that makes the kitchen feel intuitively responsive with rich tactile and olfactory details.
Winner: ChatGPT wins for a description that didn’t go too far off base while being creative and vivid.
9. Brainstorming
Prompt: “Brainstorm three book titles about finding balance in the age of AI.”
ChatGPT-5.1 generated three clear, strong, and market-ready titles that effectively capture different angles of the theme.
Claude 4.5 Sonnet offered titles with exceptional conceptual depth and provocative tension.
Winner: Claude wins for creating titles that are not just descriptive but are powerful conceptual frameworks in themselves.
Overall winner: Claude
After nine tough rounds, Claude has emerged the winner. Although it may not win every task, it wins where originality and out-of-the-box thinking really counts.
This test proves that even with the updated model, ChatGPT-5.1 does not have the imagination, emotional depth and creative understanding as Claude. So if you're staring down a blank page, that's the partner you want.
Which chatbot is your favorite for creative writing and coming up with original concepts? Let me know in the comments.
Follow Tom's Guide on Google News and add us as a preferred source to get our up-to-date news, analysis, and reviews in your feeds.
More from Tom's Guide
- ChatGPT just accidentally leaked private chats into Google Search (again) — how to stay safe
- I tried Google Photos’ new ‘Help me edit’ tool on iPhone — and I’m already obsessed
- I replaced my journal with ChatGPT for a week — here’s why I’m sticking with it
You must confirm your public display name before commenting
Please logout and then login again, you will then be prompted to enter your display name.
