Claude vs ChatGPT for writing

Every limit is published. Paid plans also have a monthly fair-use limit on AI cost: Pro $7.50, Max $20, Ultra $42, Studio $85. Using every message on your plan at typical sizes stays under it; very large messages and heavy research use it faster. Every limit, explained.

Writing: Claude Sonnet 5 passed 4 of 5 and GPT-6 Luna 5 of 5. GPT-6 Luna answered 1.9× sooner at the median, 4.5 s against 2.4 s. GPT-6 Luna cost 34.3× less, $0.0171 against $0.0005 for the 5 replies. Claude Sonnet 5's replies ran 115% longer, in tokens of reply, thinking not counted.

1. 1. Announce a second bakery shop on LinkedIn

   Everyday. Claude: 4 of 5 models passed, Claude Haiku 4.5 the cheapest at $0.0013 · GPT: 3 of 3 models passed, GPT-6 Luna the cheapest at $0.0001. The prompt in full.

   GPT-6 Luna passed this one and Claude Sonnet 5 didn't. GPT-6 Luna answered sooner (2.9 s to 4.6 s), and GPT-6 Luna's reply cost less ($0.0001 to $0.0037); Claude Sonnet 5 wrote 267 tokens of reply, GPT-6 Luna 149.

   • Claude Sonnet 5

     Failed: Graded 3.5 of 5 on average (lowest 3).

     We have some exciting news to share: Crumb & Co. is opening our second shop, this time in Leeds!
     
     From our very first bakery, we've believed in doing simple things well — proper sourdough, honest pastries, and coffee worth stopping for. Now we get to bring that same spirit to a new neighbourhood and, hopefully, a whole new group of familiar faces.
     
     Our Leeds shop opens its doors on 14 November, and we'd love for you to be there. To celebrate, every coffee on opening day is on us — no catch, just our way of saying thank you for supporting us along the way.
     
     This has been months of planning, a lot of taste-testing, and more excitement than we know what to do with. We can't wait to welcome you in, share some fresh bakes, and start building new memories together.
     
     See you there — first coffee's on the house.

     538 tokens in, 267 out · 4.6 s · $0.0037 · 1 message on Pro · answered by anthropic/claude-sonnet-5 via Claude Platform on AWS · September 27, 2026

   • GPT-6 Luna

     Passed: Graded 4.5 of 5 on average (lowest 4).

     We’re delighted to share that Crumb & Co. is opening a second shop in Leeds on 14 November!
     
     What began as one small bakery has grown thanks to the lovely people who’ve stopped by, shared a loaf, and made us feel part of the community. We’re excited to bring our freshly baked bread, pastries, and warm welcomes to a new neighbourhood.
     
     To celebrate opening day, we’ll be giving away free coffee. Pop in, say hello, and enjoy a cup with us—we’d love to meet you.
     
     Thank you to everyone who has supported Crumb & Co. so far. We can’t wait to open the doors and start this next chapter together. See you in Leeds on 14 November!

     389 tokens in, 149 out · 2.9 s · $0.0001 · 1 message on Pro · answered by openai/gpt-6-luna via OpenAI · September 27, 2026

2. 2. Rewrite corporate jargon in plain words

   Everyday. Claude: 5 of 5 models passed, Claude Haiku 4.5 the cheapest at $0.0008 · GPT: 2 of 3 models passed, GPT-6 Luna the cheapest at $0.0001. The prompt in full.

   Claude Sonnet 5 and GPT-6 Luna both passed this one. GPT-6 Luna answered sooner (2.4 s to 2.7 s), and GPT-6 Luna's reply cost less ($0.0001 to $0.0022); Claude Sonnet 5 wrote 105 tokens of reply, GPT-6 Luna 65.

   • Claude Sonnet 5

     Passed: Graded 4.7 of 5 on average (lowest 4).

     In Q3, teams worked together to focus more on customers' needs. This effort increased customer retention by 12% and reduced the rate of customer loss (churn) by 0.3 percentage points among large business clients. The company also adjusted the number of customer contact channels (like phone, email, and chat) to better manage workload and resources.

     597 tokens in, 105 out · 2.7 s · $0.0022 · 1 message on Pro · answered by anthropic/claude-sonnet-5 via Claude Platform on AWS · September 27, 2026

   • GPT-6 Luna

     Passed: Graded 4.7 of 5 on average (lowest 4).

     In Q3, teams across the company worked together to focus more on customers. Retention measures improved by 12%, and the rate of customer loss fell by 0.3 percentage points in the enterprise segment. The company also adjusted its customer contact channels to make better use of available capacity.

     417 tokens in, 124 out (59 of them reasoning) · 2.4 s · $0.0001 · 1 message on Pro · answered by openai/gpt-6-luna via OpenAI · September 27, 2026

3. 3. Decline a meeting and offer two times

   Everyday. Claude: 5 of 5 models passed, Claude Haiku 4.5 the cheapest at $0.0010 · GPT: 3 of 3 models passed, GPT-6 Luna the cheapest at $0.0001. The prompt in full.

   Claude Sonnet 5 and GPT-6 Luna both passed this one. GPT-6 Luna answered sooner (1.3 s to 2.8 s), and GPT-6 Luna's reply cost less ($0.0001 to $0.0026); Claude Sonnet 5 wrote 149 tokens of reply, GPT-6 Luna 64.

   • Claude Sonnet 5

     Passed: Graded 5.0 of 5 on average (lowest 5).

     Subject: Re: Vendor Demo – Wednesday 11:00
     
     Hi Tom,
     
     Thanks for the invite! Unfortunately, I have a board presentation at that time on Wednesday, so I won't be able to join.
     
     Would Thursday at 10:00 or Friday at 14:00 work instead? Happy to make either one work on my end.
     
     Let me know what suits you, and I'll get it on the calendar.
     
     Best,
     [Your Name]

     542 tokens in, 149 out · 2.8 s · $0.0026 · 1 message on Pro · answered by anthropic/claude-sonnet-5 via Claude Platform on AWS · September 27, 2026

   • GPT-6 Luna

     Passed: Graded 5.0 of 5 on average (lowest 5).

     Subject: Vendor Demo
     
     Hi Tom,
     
     Thanks for the invitation. I’m presenting to the board on Wednesday at 11:00, so I can’t make the demo then. Would Thursday at 10:00 or Friday at 14:00 work instead?
     
     Best,  
     [Your Name]

     395 tokens in, 64 out · 1.3 s · $0.0001 · 1 message on Pro · answered by openai/gpt-6-luna via OpenAI · September 27, 2026

4. 4. A product announcement with five rules

   Hard. Claude: 3 of 5 models passed, Claude Sonnet 5.5 the cheapest at $0.0030 · GPT: 3 of 3 models passed, GPT-6 Luna the cheapest at $0.0001. The prompt in full.

   Claude Sonnet 5 and GPT-6 Luna both passed this one. GPT-6 Luna answered sooner (1.5 s to 4.5 s), and GPT-6 Luna's reply cost less ($0.0001 to $0.0035); Claude Sonnet 5 wrote 235 tokens of reply, GPT-6 Luna 73.

   • Claude Sonnet 5

     Passed: Graded 4.3 of 5 on average (lowest 4).

     **Introducing the Nordvik 2 Desk Lamp**
     
     Meet the Nordvik 2, thoughtfully redesigned for how you actually work. Five-step dimming lets you fine-tune brightness for any task, from focused reading to ambient evening light. The 270-degree rotating arm adjusts to virtually any angle, giving you precise control over where light falls. USB-C charging means one cable does it all, keeping your desk clutter-free and compatible with everything else you own.
     
     Priced at €79, the Nordvik 2 brings premium design and functionality within easy reach. It begins shipping on 1 December, just in time to brighten your winter workspace.
     
     Ready to see how the right light changes the way you work?

     554 tokens in, 235 out · 4.5 s · $0.0035 · 1 message on Pro · answered by anthropic/claude-sonnet-5 via Claude Platform on AWS · September 27, 2026

   • GPT-6 Luna

     Passed: Graded 4.0 of 5 on average (lowest 3).

     Meet the Nordvik 2, a desk lamp designed to bring flexible light to your workspace. Choose from five brightness levels, charge it via USB-C, and turn its arm up to 270 degrees to direct light where you need it. The Nordvik 2 costs €79 and ships on 1 December. Ready to brighten your desk?

     406 tokens in, 73 out · 1.5 s · $0.0001 · 1 message on Pro · answered by openai/gpt-6-luna via OpenAI · September 27, 2026

5. 5. Argue both sides of free buses

   Hard. Claude: 2 of 5 models passed, Claude Haiku 4.5 the cheapest at $0.0014 · GPT: 3 of 3 models passed, GPT-6 Luna the cheapest at $0.0001. The prompt in full.

   Claude Sonnet 5 and GPT-6 Luna both passed this one. GPT-6 Luna answered sooner (2.5 s to 6.6 s), and GPT-6 Luna's reply cost less ($0.0001 to $0.0050); Claude Sonnet 5 wrote 401 tokens of reply, GPT-6 Luna 186.

   • Claude Sonnet 5

     Passed: Graded 4.0 of 5 on average (lowest 4).

     Free buses would boost ridership, cut traffic congestion, and reduce emissions by encouraging drivers to switch to transit. Lower-income residents, who rely most heavily on public transportation, would benefit enormously from eliminated fare burdens, improving access to jobs, education, and healthcare. Free transit also simplifies operations, speeding boarding times and eliminating costly fare collection infrastructure. Cities that have tried this, like Kansas City, saw ridership increases and stronger community mobility, suggesting free buses are both an equity tool and a practical urban planning solution.
     
     However, fares provide crucial revenue that funds maintenance, expansion, and service reliability. Eliminating them shifts costs entirely onto taxpayers, straining budgets already stretched thin, potentially forcing service cuts elsewhere. Free buses often attract overcrowding without corresponding investment in more vehicles, degrading service quality for existing riders. Critics also note that the poorest riders may already receive subsidized passes, making universal free fares a regressive policy that primarily benefits wealthier occasional riders. Better targeted subsidies, they argue, would achieve equity goals more efficiently than blanket free fares citywide.

     518 tokens in, 401 out · 6.6 s · $0.0050 · 1 message on Pro · answered by anthropic/claude-sonnet-5 via Claude Platform on AWS · September 27, 2026

   • GPT-6 Luna

     Passed: Graded 4.3 of 5 on average (lowest 4).

     Making buses free would give every resident reliable access to jobs, school, health care, and daily necessities, regardless of income. It would also speed boarding, reduce fare-enforcement costs, and encourage people to leave cars at home—easing congestion, pollution, and parking pressure. A city that treats public transit as essential infrastructure, rather than a service people must qualify to afford, can build a more connected and equitable community.
     
     But “free” buses still have to be paid for, and the bill could divert money from more urgent improvements: frequent service, safer stops, cleaner vehicles, and routes to underserved neighborhoods. If buses remain slow, crowded, or unreliable, removing fares may do little to persuade drivers to switch, while reducing funds that could make transit genuinely useful. Targeted fare assistance can protect riders who need it most without subsidizing every trip, including those by people who can readily afford to pay.

     384 tokens in, 186 out · 2.5 s · $0.0001 · 1 message on Pro · answered by openai/gpt-6-luna via OpenAI · September 27, 2026

• Claude Fable 5.1 vs GPT-6 Astra: Claude Fable 5.1 3 of 5, GPT-6 Astra 5; GPT-6 Astra answered 1.2× sooner at the median and cost 1.6× less. Claude Fable 5.1 vs GPT-6 Astra, on every job.

• Claude Fable 5.1 vs GPT-6 Sol: Claude Fable 5.1 3 of 5, GPT-6 Sol 4; GPT-6 Sol answered 1.7× sooner at the median and cost 7.6× less.

• Claude Fable 5.1 vs GPT-6 Luna: Claude Fable 5.1 3 of 5, GPT-6 Luna 5; GPT-6 Luna answered 2.5× sooner at the median and cost 173.5× less.

• Claude Opus 5.5 vs GPT-6 Astra: Claude Opus 5.5 4 of 5, GPT-6 Astra 5; GPT-6 Astra answered 1.7× sooner at the median and cost 1.7× less. Claude Opus 5.5 vs GPT-6 Astra, on every job.

• Claude Opus 5.5 vs GPT-6 Sol: 4 of 5 each; GPT-6 Sol answered 2.4× sooner at the median and cost 8.2× less. Claude Opus 5.5 vs GPT-6 Sol, on every job.

• Claude Opus 5.5 vs GPT-6 Luna: Claude Opus 5.5 4 of 5, GPT-6 Luna 5; GPT-6 Luna answered 3.5× sooner at the median and cost 185.7× less.

• Claude Sonnet 5.5 vs GPT-6 Astra: Claude Sonnet 5.5 4 of 5, GPT-6 Astra 5; Claude Sonnet 5.5 answered 2.0× sooner at the median and cost 3.3× less.

• Claude Sonnet 5.5 vs GPT-6 Sol: 4 of 5 each; Claude Sonnet 5.5 answered 1.5× sooner at the median and GPT-6 Sol cost 1.5× less. Claude Sonnet 5.5 vs GPT-6 Sol, on every job.

• Claude Sonnet 5.5 vs GPT-6 Luna: Claude Sonnet 5.5 4 of 5, GPT-6 Luna 5; they took about as long and GPT-6 Luna cost 34.3× less.

• Claude Sonnet 5 vs GPT-6 Astra: Claude Sonnet 5 4 of 5, GPT-6 Astra 5; they took about as long and Claude Sonnet 5 cost 3.3× less.

• Claude Sonnet 5 vs GPT-6 Sol: 4 of 5 each; GPT-6 Sol answered 1.3× sooner at the median and cost 1.5× less. Claude Sonnet 5 vs GPT-6 Sol, on every job.

• Claude Sonnet 5 vs GPT-6 Luna: Claude Sonnet 5 4 of 5, GPT-6 Luna 5; GPT-6 Luna answered 1.9× sooner at the median and cost 34.3× less.

• Claude Haiku 4.5 vs GPT-6 Astra: Claude Haiku 4.5 4 of 5, GPT-6 Astra 5; Claude Haiku 4.5 answered 2.2× sooner at the median and cost 9.8× less.

• Claude Haiku 4.5 vs GPT-6 Sol: 4 of 5 each; Claude Haiku 4.5 answered 1.6× sooner at the median and cost 2.0× less. Claude Haiku 4.5 vs GPT-6 Sol, on every job.

• Claude Haiku 4.5 vs GPT-6 Luna: Claude Haiku 4.5 4 of 5, GPT-6 Luna 5; they took about as long and GPT-6 Luna cost 11.4× less. Claude Haiku 4.5 vs GPT-6 Luna, on every job.

Every Claude and GPT reply above was checked the same way as every other model's, by the rules published with the prompts: how the writing prompts are scored, and each one in full.

• The lineups

  Claude: 5 models, Claude Fable 5.1, Claude Opus 5.5, Claude Sonnet 5.5, Claude Sonnet 5, and Claude Haiku 4.5. GPT: 3 models, GPT-6 Astra, GPT-6 Sol, and GPT-6 Luna.

• Price per message

  The least expensive Claude model is Claude Haiku 4.5 (250 messages a month on Pro); the least expensive GPT model is GPT-6 Luna (60 messages a day on Pro).

• Context window

  Claude goes up to 1M tokens (Claude Fable 5.1); GPT up to 1.05M tokens (GPT-6 Astra).

• Images and PDFs

  Every model here reads images. Every model here takes a PDF as a whole file.

• On the Free plan

  Free's one-time trial of 5 messages covers Claude Sonnet 5.5, Claude Sonnet 5, Claude Haiku 4.5, GPT-6 Sol, and GPT-6 Luna. Paid plans have every model, with messages every month.

In llmwise, a message to Claude or GPT goes to the model's maker, or through OpenRouter when llmwise can't reach the maker directly. The Privacy Policy has the details.