ChatGPT vs DeepSeek 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: GPT-6 Luna passed 5 of 5 and DeepSeek V4.1 Flash 4 of 5. GPT-6 Luna answered 1.3× sooner at the median, 2.4 s against 3.1 s. GPT-6 Luna cost 2.9× less, $0.0005 against $0.0014 for the 5 replies. DeepSeek V4.1 Flash's replies ran 17% longer, in tokens of reply, thinking not counted.

1. 1. Announce a second bakery shop on LinkedIn

   Everyday. GPT: 3 of 3 models passed, GPT-6 Luna the cheapest at $0.0001 · DeepSeek: 2 of 2 models passed, DeepSeek V4.1 Flash the cheapest at $0.0002. The prompt in full.

   GPT-6 Luna and DeepSeek V4.1 Flash both passed this one. They took about as long (GPT-6 Luna 2.9 s, DeepSeek V4.1 Flash 2.9 s), and GPT-6 Luna's reply cost less ($0.0001 to $0.0002); GPT-6 Luna wrote 149 tokens of reply, DeepSeek V4.1 Flash 184.

   • 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

   • DeepSeek V4.1 Flash

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

     Some news we've been sitting on for a while: Crumb & Co. is opening a second shop, in Leeds, on 14 November.
     
     We started six years ago with one oven, one counter and a lot of very early mornings. The idea was simple — good bread, made slowly, sold to people who'd walk a bit out of their way for it. That we get to do it in a second place still feels a little unreal.
     
     The new shop will have the same loaves, the same pastries, the same slightly chaotic Saturday mornings. There's more seating, and a window that actually gets sun.
     
     To say thank you, coffee is free all day on opening day. No catch, no app, no code — just come in and we'll make you one.
     
     We'd love to see you there. Thank you for getting us this far.
     
     — The Crumb & Co. team

     428 tokens in, 203 out (19 of them reasoning) · 2.9 s · $0.0002 · 1 message on Pro · answered by deepseek/deepseek-v4.1-flash via GMICloud · September 27, 2026

2. 2. Rewrite corporate jargon in plain words

   Everyday. GPT: 2 of 3 models passed, GPT-6 Luna the cheapest at $0.0001 · DeepSeek: 1 of 2 models passed, DeepSeek V4 Pro at $0.0006. The prompt in full.

   GPT-6 Luna passed this one and DeepSeek V4.1 Flash didn't. DeepSeek V4.1 Flash answered sooner (1.2 s to 2.4 s), and GPT-6 Luna's reply cost less ($0.0001 to $0.0003); GPT-6 Luna wrote 65 tokens of reply, DeepSeek V4.1 Flash 68.

   • 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

   • DeepSeek V4.1 Flash

     Failed: Graded 3.7 of 5 on average (lowest 3).

     In the third quarter, teams from across the company worked together on a project that put customers first. For our large business customers, retention rose 12%, and churn fell by 30 basis points. We also adjusted our channels — online, in-store, and other ways customers reach us — to use our resources more efficiently.

     459 tokens in, 129 out (61 of them reasoning) · 1.2 s · $0.0003 · 1 message on Pro · answered by deepseek/deepseek-v4.1-flash via Parasail · September 27, 2026

3. 3. Decline a meeting and offer two times

   Everyday. GPT: 3 of 3 models passed, GPT-6 Luna the cheapest at $0.0001 · DeepSeek: 2 of 2 models passed, DeepSeek V4.1 Flash the cheapest at $0.0002. The prompt in full.

   GPT-6 Luna and DeepSeek V4.1 Flash both passed this one. GPT-6 Luna answered sooner (1.3 s to 3.1 s), and GPT-6 Luna's reply cost less ($0.0001 to $0.0002); GPT-6 Luna wrote 64 tokens of reply, DeepSeek V4.1 Flash 127.

   • 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

   • DeepSeek V4.1 Flash

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

     **Subject:** Re: Vendor demo — Wednesday 11:00
     
     Hi Tom,
     
     Thanks for the invitation to the vendor demo. Unfortunately, I can't make Wednesday at 11:00 — I'm presenting to the board at that time.
     
     I'd still like to join, though. Would either of these work instead?
     
     - Thursday at 10:00
     - Friday at 14:00
     
     Let me know which suits you, and I'll confirm. If neither works, feel free to send notes or a recording and I'll follow up.
     
     Thanks again for thinking of me.
     
     Best regards,
     [Your name]

     434 tokens in, 157 out (30 of them reasoning) · 3.1 s · $0.0002 · 1 message on Pro · answered by deepseek/deepseek-v4.1-flash via Parasail · September 27, 2026

4. 4. A product announcement with five rules

   Hard. GPT: 3 of 3 models passed, GPT-6 Luna the cheapest at $0.0001 · DeepSeek: 1 of 2 models passed, DeepSeek V4.1 Flash at $0.0004. The prompt in full.

   GPT-6 Luna and DeepSeek V4.1 Flash both passed this one. GPT-6 Luna answered sooner (1.5 s to 6.2 s), and GPT-6 Luna's reply cost less ($0.0001 to $0.0004); GPT-6 Luna wrote 73 tokens of reply, DeepSeek V4.1 Flash 91.

   • 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

   • DeepSeek V4.1 Flash

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

     Introducing the Nordvik 2, a desk lamp built for the way you actually work. Five dimming steps take you from bright task light to a soft evening glow. The arm turns a full 270 degrees, so you can aim light exactly where it is needed. Recharging is simple, via USB-C. The Nordvik 2 costs €79 and ships on 1 December. Would you like to see how it looks on your desk?

     445 tokens in, 495 out (404 of them reasoning) · 6.2 s · $0.0004 · 1 message on Pro · answered by deepseek/deepseek-v4.1-flash via CoreWeave · September 27, 2026

5. 5. Argue both sides of free buses

   Hard. GPT: 3 of 3 models passed, GPT-6 Luna the cheapest at $0.0001 · DeepSeek: 1 of 2 models passed, DeepSeek V4.1 Flash at $0.0003. The prompt in full.

   GPT-6 Luna and DeepSeek V4.1 Flash both passed this one. GPT-6 Luna answered sooner (2.5 s to 3.9 s), and GPT-6 Luna's reply cost less ($0.0001 to $0.0003); GPT-6 Luna wrote 186 tokens of reply, DeepSeek V4.1 Flash 160.

   • 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

   • DeepSeek V4.1 Flash

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

     Free buses work. They cut car trips, congestion, and carbon, and they help the riders who need them most. Boarding speeds up, so routes run faster and cheaper to operate. The cost is modest—often less than a parking garage or a lane widening—and downtown shops gain customers. A city that charges nothing at the farebox treats mobility as a right, not a fee.
     
     But somebody still pays. Fares cover a real slice of operating costs, and replacing them means higher taxes or cuts elsewhere—often taxes that hit the poor hardest. Most riders already pay, so the subsidy flows to everyone, including the comfortable. Crowding slows buses, driving away choice riders. Better to means-test discounts, keep service frequent, and spend scarce dollars where they move the most people.

     423 tokens in, 282 out (122 of them reasoning) · 3.9 s · $0.0003 · 1 message on Pro · answered by deepseek/deepseek-v4.1-flash via StreamLake · September 27, 2026

• GPT-6 Astra vs DeepSeek V4 Pro: GPT-6 Astra 5 of 5, DeepSeek V4 Pro 3; DeepSeek V4 Pro answered 2.1× sooner at the median and cost 24.9× less.

• GPT-6 Astra vs DeepSeek V4.1 Flash: GPT-6 Astra 5 of 5, DeepSeek V4.1 Flash 4; DeepSeek V4.1 Flash answered 1.6× sooner at the median and cost 39.0× less.

• GPT-6 Sol vs DeepSeek V4 Pro: GPT-6 Sol 4 of 5, DeepSeek V4 Pro 3; DeepSeek V4 Pro answered 1.5× sooner at the median and cost 5.1× less. GPT-6 Sol vs DeepSeek V4 Pro, on every job.

• GPT-6 Sol vs DeepSeek V4.1 Flash: 4 of 5 each; DeepSeek V4.1 Flash answered 1.1× sooner at the median and cost 7.9× less.

• GPT-6 Luna vs DeepSeek V4 Pro: GPT-6 Luna 5 of 5, DeepSeek V4 Pro 3; they took about as long and GPT-6 Luna cost 4.5× less. GPT-6 Luna vs DeepSeek V4 Pro, on every job.

• GPT-6 Luna vs DeepSeek V4.1 Flash: GPT-6 Luna 5 of 5, DeepSeek V4.1 Flash 4; GPT-6 Luna answered 1.3× sooner at the median and cost 2.9× less. GPT-6 Luna vs DeepSeek V4.1 Flash, on every job.

Every GPT and DeepSeek 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

  GPT: 3 models, GPT-6 Astra, GPT-6 Sol, and GPT-6 Luna. DeepSeek: 2 models, DeepSeek V4 Pro and DeepSeek V4.1 Flash.

• Price per message

  The least expensive GPT model is GPT-6 Luna (60 messages a day on Pro); the least expensive DeepSeek model is DeepSeek V4.1 Flash (60 messages a day on Pro).

• Context window

  GPT goes up to 1.05M tokens (GPT-6 Astra); DeepSeek up to 1.05M tokens (DeepSeek V4 Pro).

• Images and PDFs

  DeepSeek V4 Pro doesn't read images. DeepSeek V4 Pro and DeepSeek V4.1 Flash get a PDF's text rather than the file itself.

• On the Free plan

  Free's one-time trial of 5 messages covers GPT-6 Sol, GPT-6 Luna, DeepSeek V4 Pro, and DeepSeek V4.1 Flash. Paid plans have every model, with messages every month.

In llmwise, a message to GPT goes to its maker, OpenAI, or through OpenRouter when llmwise can't reach the maker directly. DeepSeek models are served only through OpenRouter, by endpoints that don't store or train on prompts. The Privacy Policy has the details.