It almost seems as if a new chatbot launches every other week, or at the very least, new features are added to current ones.
With the recent launch of Anthropic’s Claude Haiku 5.5, a smaller, cheaper model designed for speed and efficiency, I had to know how it stands up against ChatGPT-6 for every day prompts. I have already compared ChatGPT-6 to Claude’s sister model Opus 5.5, and couldn’t help but wonder how the lighter model compared, too.
OpenAI’s GPT-6 powers the latest generation of ChatGPT, with an emphasis on more useful answers and interactive tools. Both promise practical help, but I wanted to see how well they handled my specific requests.
I used identical prompts to test meal planning, creative writing, shopping research, family scheduling and building a household budget calculator. I judged the responses on how closely they followed my instructions and how much work they left for me to do. Here’s what happened.
1. The common-sense reasoning test
Prompt: I’m hosting Thanksgiving dinner for 10 people. One is vegan, one has a nut allergy, two are picky kids and I only have one oven. Create a menu and cooking schedule that gets everything on the table at 5 p.m. without spending more than $175.
Claude accurately calculated the turkey cooking time well, including a generous resting window before a 5:00 p.m dinner. It offered smart equipment hacks, kid friendly options, and allergy suggestions.
ChatGPT made almost all side dishes vegan by default (mashed potatoes, stuffing, green beans, carrots, gravy) using olive oil and dairy-free milk. This eliminates the need to cook separate vegan/non-vegan batches for every side item. It offered a realistic budget beakdown and a backup plan.
Winner: ChatGPT wins. The single hardest part of cooking Thanksgiving for guests with dietary restrictions in a one-oven kitchen is juggling multiple dishes. ChatGPT’s strategy of making all side dishes vegan by default reduces cooking complexity by 50%. It was efficient and helpful and the images were a nice added touch.
2. The creative writing test
Prompt: Write the opening 250 words of a novel about a young boy who stumbles upon a cave and discovers a kind dragon inside who befriends and protects him from bullies
Claude offered excellent atmospheric and environmental details. It also offered distinct dragon personality details.
ChatGPT Immediately launched into the middle of the action—Oliver is actively running away from bullies in sentence one. The story included interesting sensory details and fun fantasy elements.
Winner: ChatGPT wins for following the prompt the best and for superior middle-grade tone.
3. The research and fact-checking test
Prompt: Find three pairs of top rated running shoes that are under $150, and perfect for a decades long runner with a high arch running 35 miles a week
Claude accurately explained why high-arched runners logging high mileage (35 miles/week) need neutral, high-cushion trainers over stability shoes (extra shock absorption vs. preventing overpronation as well as including a top three list of running shoes.
ChatGPT delivered real-world data to back up the running shoe suggestions and ended with personal follow up questions to help me find the perfect pair.
Winner: Claude wins. The chatbot offered superior podiatric insight while identifying the mechanics of high arches (rigidity and poor shock absorption) and correctly steered me toward plush, neutral max-cushion models (Glycerin, Clifton, Ride). ChatGPT recommended the Brooks Ghost, which is famously firm and less ideal for high arches requiring shock protection over 35 weekly miles.
4. The chaotic schedule test
Prompt: I’m a working parent with three kids. The bus drops the kids off at 4:25, soccer starts at 6:30, and I need to swing by the post office before it closes at 6pm. Plus, dinner needs to be ready by 6:30. I also have a work deadline at 4. Build me a realistic afternoon schedule, identify impossible conflicts and suggest what I should delegate or reschedule.
Claude offered practical ways to get through the afternoon, with suggestions for what to delegate or reschedule that felt realistic for a busy parent.
ChatGPT addressed the dinner and soccer clash and built in buffers to keep one delay from derailing the schedule. It also included a table outlining potential problems and how to handle them.
Winner: Claude. Its advice felt more personal and easier to act on, with clever parenting shortcuts that reduced the stress of an already packed afternoon.
5. The build-something-useful test
Prompt: Build me a working household budget calculator. I want to enter monthly income, fixed bills, groceries, subscriptions and savings goals. Show how much money remains and let me change amounts to see the impact immediately. Make it simple enough for someone who has never written code to use.
Claude built a straightforward budget calculator with sample bills already entered and room to add my own expenses.
ChatGPT created a more polished app with editable amounts I could tailor to my household. I could also download it to use on the go or offline.
Winner: ChatGPT. Its calculator was easier to customize and offered useful extras, including a reset option and mobile access, that made it something I’d keep using.
Overall winner: ChatGPT
ChatGPT won this face-off 3–2, taking the Thanksgiving planning, creative writing and budget calculator rounds. Its best responses made my next step easier, whether that meant cooking one set of sides for everyone or giving me a budget tool I could immediately adjust.
Claude Haiku 5.5 earned its two wins with advice that felt more tailored to my situation. Its running shoe response better addressed the needs I described, while its afternoon schedule offered direct, practical ways to manage competing demands.
There’s no doubt that Haiku 5.5 is a good model for everyday questions, but for my mix of tasks, ChatGPT would probably be my first choice based on these tests.
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. Subscribe to Tom’s Guide on YouTube and follow us on TikTok.
More from Tom’s Guide
Next Badge:
More Comments/Likes Until Your Next Badge