DeepSeek V4 Pro · Writing
DeepSeek V4 Pro for writing: our test runs
DeepSeek V4 Pro ran our 5 writing prompts in llmwise, through the same pipeline your messages take. Here's every reply as it came, whether it passed, what it cost and what it counts as on Pro, next to the other models.
Based on 5 of our test runs on , through OpenRouter with the app's own prompt and settings.
Short answer
In our test runs on September 27, 2026, DeepSeek V4 Pro passed 3 of 5 of our writing prompts (neither hard one): tied 11th of 15 models. Each reply counted as 1 message on Pro, where it has 250 a month, and cost $0.0004 to run on average.
Prompt by prompt
Each of the 5 writing prompts, and how DeepSeek V4 Pro's reply did.
| Prompt | Result | Cost | Time |
|---|---|---|---|
| Announce a second bakery shop on LinkedInEveryday | Passed: Graded 4.3 of 5 on average (lowest 4). | $0.0008 | 1.9 s |
| Rewrite corporate jargon in plain wordsEveryday | Passed: Graded 4.0 of 5 on average (lowest 3). | $0.0006 | 1.2 s |
| Decline a meeting and offer two timesEveryday | Passed: Graded 5.0 of 5 on average (lowest 5). | $0.0003 | 2.2 s |
| A product announcement with five rulesHard | Failed: Graded 3.7 of 5 on average (lowest 2); but doesn't end with a question. | $0.0001 | 3.8 s |
| Argue both sides of free busesHard | Failed: Graded 4.0 of 5 on average (lowest 4); but a paragraph of 95 words, over the 90 allowed. | $0.0003 | 7.8 s |
Every writing reply from DeepSeek V4 Pro
Announce a second bakery shop on LinkedIn
An everyday prompt: “Announce a second bakery shop on LinkedIn”, in full.
Passed: Graded 4.3 of 5 on average (lowest 4).
Some lovely news from Crumb & Co. This November, we're opening the doors to our second bakery, this time in Leeds. It's been a long-held dream to bring our bread, pastries and cakes to a new neighbourhood, and we're so pleased it's finally happening. From 14 November, you'll find us on the corner of Briggate and Boar Lane, with the same slow-fermented sourdough, buttery croissants and cakes we bake each morning. To celebrate, we'd love to buy you a coffee on opening day. Drop by on Thursday 14 November and your first cup is on us, whether you're stopping for a pastry or just saying hello. If you're in Leeds, we'd love to see you. Thank you to everyone who has supported us so far. We couldn't have done this without you.
400 tokens in, 213 out (37 of them reasoning) · 1.9 s · $0.0008 · 1 message on Pro · answered by deepseek/deepseek-v4-pro-0813 via Wafer ·
Rewrite corporate jargon in plain words
An everyday prompt: “Rewrite corporate jargon in plain words”, in full.
Passed: Graded 4.0 of 5 on average (lowest 3).
In the third quarter, our teams worked together to put customers first. As a result, retention among enterprise customers rose by 12%, and churn fell by 0.30 percentage points. We also adjusted our support and communication channels to use resources more efficiently.
431 tokens in, 139 out (85 of them reasoning) · 1.2 s · $0.0006 · 1 message on Pro · answered by deepseek/deepseek-v4-pro-0813 via Wafer ·
Decline a meeting and offer two times
An everyday prompt: “Decline a meeting and offer two times”, in full.
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 on Wednesday at 11:00. Unfortunately, I won't be able to make it — I'm presenting to the board at that time. Would Thursday at 10:00 or Friday at 14:00 work instead? Either would suit me well. Best, [Your name]
406 tokens in, 140 out (57 of them reasoning) · 2.2 s · $0.0003 · 1 message on Pro · answered by deepseek/deepseek-v4-pro-0813 via Ionstream ·
A product announcement with five rules
A hard prompt: “A product announcement with five rules”, in full.
Failed: Graded 3.7 of 5 on average (lowest 2); but doesn't end with a question.
Introducing the Nordvik 2 desk lamp. It dims in five steps, so you can tune the light from bright task work to a soft evening glow. It charges over USB-C, using the same cable you already carry. The arm turns 270 degrees, giving you easy control over where the light falls. The Nordvik 2 costs €79 and ships on 1 December, in time for the darker months. A desk lamp that adapts this cleanly to how you work deserves a closer look, doesn't it.
418 tokens in, 141 out (31 of them reasoning) · 3.8 s · $0.0001 · 1 message on Pro · answered by deepseek/deepseek-v4-pro-0813 via Baidu ·
Argue both sides of free buses
A hard prompt: “Argue both sides of free buses”, in full.
Failed: Graded 4.0 of 5 on average (lowest 4); but a paragraph of 95 words, over the 90 allowed.
Making buses free turns them into a genuine public good rather than a service you must pay to use. It removes the fare barrier that keeps low-income riders from jobs, medical care, and school, while also speeding up boarding since no one fumbles for change. Fewer cars on the road cuts congestion, pollution, and traffic deaths, and a fare-free system is cheaper to run because you eliminate fare collection machines, enforcement, and revenue handling. A city that values mobility and fairness should treat transit like libraries and sidewalks: open to everyone, paid for by everyone. Eliminating fares is a poor trade, however, because it spends scarce transit dollars on people who can already afford to ride while doing nothing to improve service for those who cannot. Free buses attract crowding and rowdiness, deterring paying-quality riders and raising security costs. The revenue loss forces cuts to routes and frequency, which hurts the very riders the policy claims to help. A better approach is means-tested fare discounts and fare capping, which target subsidies precisely while preserving the funding needed for reliable, frequent, and expanding service.
395 tokens in, 263 out (41 of them reasoning) · 7.8 s · $0.0003 · 1 message on Pro · answered by deepseek/deepseek-v4-pro-0813 via StreamLake ·
How it compares on writing
DeepSeek V4 Pro beside the job's picks among every model. 2 models share the hard-problems pick, and DeepSeek V4 Pro isn't one of them.
| Model | Passed | On Pro | Cost per reply |
|---|---|---|---|
| GLM 5.3Our pick: best value | 4 of 5 | 250 a month on Pro | $0.0006 |
| GPT-6 LunaOur pick: hard problems (shared), everyday | 5 of 5 | 60 a day on Pro | $0.0001 |
| DeepSeek V4 Pro | 3 of 5 | 250 a month on Pro | $0.0004 |
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.
How these runs were done
Rubric (graded). The grader model scores the reply from 1 to 5 on each published criterion. It passes with an average of 4 or more and no criterion under 3, and only if it also meets the prompt's automatic rules (length, words it must or mustn't use).
The grader is Claude Opus 5.5 at low reasoning effort; its own replies are graded by GPT-6 Astra, so no model grades itself. Its prompt and every rubric are on the methods page.
DeepSeek V4 Pro, and writing, elsewhere
- DeepSeek V4 Pro: price, limits and messages on every plan
- The best AI for writing: our picks among every model
- DeepSeek for writing
- DeepSeek V4 Pro for coding: our test runs
- DeepSeek V4 Pro for math: our test runs
- DeepSeek V4 Pro for data analysis: our test runs
- Claude Opus 5.5 vs DeepSeek V4 Pro
- Rewrite for clarity: the prompt, and its price on each model
- Our test runs: 50 prompts on every model
Questions
Is DeepSeek V4 Pro good for writing?
In our test runs it passed 3 of 5 writing prompts, tied 11th of the 15 models in llmwise. Every reply is on this page, so you can judge them yourself.
How many of my messages does a writing reply from DeepSeek V4 Pro use?
1 message each on Pro, where it has 250 a month on Pro. The price of a message is fixed and shown before you send it, however long the reply.
How were these runs done?
The same way for every model: each prompt sent through llmwise's own pipeline, each reply checked the same way. The methods page has every prompt and how each is scored.
Claude, GPT, Gemini, DeepSeek, Grok, Kimi, and GLM, in one chat.
See what a message costs before you send it. Free is 5 messages to try; sign in with an email link, no password or card.