Skip to content

Kimi K3.1 Pro · Writing

Kimi K3.1 Pro for writing: our test runs

Kimi K3.1 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 November 17, 2026, Kimi K3.1 Pro passed 5 of 5 of our writing prompts (both hard ones): tied 1st of 27 models. Each reply counted as 1 message on Pro, where it has 125 a month, and cost $0.0050 to run on average.

Prompt by prompt

Each of the 5 writing prompts, and how Kimi K3.1 Pro's reply did.

Kimi K3.1 Pro on each writing prompt
PromptResultCostTime
Announce a second bakery shop on LinkedInEverydayPassed: Graded 4.5 of 5 on average (lowest 4).$0.00546.9 s
Rewrite corporate jargon in plain wordsEverydayPassed: Graded 4.7 of 5 on average (lowest 4).$0.00527.0 s
Decline a meeting and offer two timesEverydayPassed: Graded 5.0 of 5 on average (lowest 5).$0.00313.7 s
A product announcement with five rulesHardPassed: Graded 4.7 of 5 on average (lowest 4).$0.00444.7 s
Argue both sides of free busesHardPassed: Graded 4.0 of 5 on average (lowest 4).$0.006812.3 s

Every writing reply from Kimi K3.1 Pro

  1. Announce a second bakery shop on LinkedIn

    An everyday prompt: “Announce a second bakery shop on LinkedIn”, in full.

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

    We’ve got some lovely news to share: Crumb & Co. is opening a second shop in Leeds on 14 November.
    
    As a small bakery, every new chapter means a great deal to us. This one is possible because of the customers, neighbours and friends who have supported us, bought our bread, shared our cakes and cheered us on. Thank you.
    
    Our new shop will bring the same fresh bakes, friendly service and care that we put into everything we make. We can’t wait to meet more of the local community and become part of your everyday routines, whether you’re picking up a morning pastry or treating yourself to something sweet.
    
    On opening day, coffee is on us. Come in, enjoy a free coffee and help us celebrate.
    
    We look forward to welcoming you to Crumb & Co. Leeds on 14 November.

    654 tokens in, 411 out (228 of them reasoning) · 6.9 s · $0.0054 · 1 message on Pro · answered by moonshotai/kimi-k3.1-pro via Together ·

  2. Rewrite corporate jargon in plain words

    An everyday prompt: “Rewrite corporate jargon in plain words”, in full.

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

    In Q3, our teams worked together to put customers at the center of our work. The initiative improved customer retention by 12% and reduced the rate at which enterprise customers leave by 30 basis points (0.3 percentage points). We also adjusted our customer contact channels to make better use of our team’s capacity.

    682 tokens in, 387 out (309 of them reasoning) · 7.0 s · $0.0052 · 1 message on Pro · answered by moonshotai/kimi-k3.1-pro via Together ·

  3. 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: Rescheduling Wednesday’s Vendor Demo**
    
    Hi Tom,
    
    Thanks for the invitation to the vendor demo on Wednesday at 11:00. Unfortunately, I’ll be presenting to the board at that time and won’t be able to attend.
    
    Could we reschedule for Thursday at 10:00 or Friday at 14:00 instead? Either time works for me.
    
    Sorry for any inconvenience, and thanks for your flexibility.
    
    Best,  
    [Your Name]

    660 tokens in, 177 out (71 of them reasoning) · 3.7 s · $0.0031 · 1 message on Pro · answered by moonshotai/kimi-k3.1-pro via Together ·

  4. A product announcement with five rules

    A hard prompt: “A product announcement with five rules”, in full.

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

    Meet the Nordvik 2 desk lamp, designed to bring adaptable light to desks, bedside tables, and reading corners. Its arm turns 270 degrees, while five-step dimming lets you move from focused task lighting to a softer evening glow. USB-C charging keeps power simple and convenient, whether you are setting up a permanent workspace or moving the lamp around your home. The Nordvik 2 costs €79 and ships on 1 December. Clean lines, flexible positioning, and practical controls make it an easy addition to your daily routine. Where will you place yours?

    671 tokens in, 305 out (178 of them reasoning) · 4.7 s · $0.0044 · 1 message on Pro · answered by moonshotai/kimi-k3.1-pro via Together ·

  5. Argue both sides of free buses

    A hard prompt: “Argue both sides of free buses”, in full.

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

    Making buses free would turn public transit into genuine public infrastructure: simple to use, available in a crisis, and no longer a daily calculation for households choosing between groceries and a ride. Ridership would rise, cars would leave crowded streets, and low-wage workers, students, and seniors would gain access to jobs, schools, and care. Eliminating fares also cuts enforcement costs, speeds boarding, and replaces a regressive charge with broader funding. A city that subsidizes roads should make its shared buses equally easy to board.
    
    Yet free buses can consume transit budgets without delivering better transit. When fares disappear, agencies may cut frequency, maintenance, and route expansions—the things riders value most; a free bus that arrives late or overcrowded is not meaningful access. Crowding could worsen, while people taking unnecessary short trips gain benefits without reducing pollution. Limited public money might help low-income riders more through targeted passes and help everyone through reliable service. The goal is affordable, useful mobility, not zero fares at any cost.

    649 tokens in, 550 out (334 of them reasoning) · 12.3 s · $0.0068 · 1 message on Pro · answered by moonshotai/kimi-k3.1-pro via Together ·

How it compares on writing

Kimi K3.1 Pro beside the job's picks among every model. 14 models share the hard-problems pick, Kimi K3.1 Pro among them.

Kimi K3.1 Pro beside other models on writing
ModelPassedOn ProCost per reply
Gemini 3.8 FlashOur pick: hard problems (shared), best value5 of 5250 a month on Pro$0.0005
DeepSeek V4.1 FlashOur pick: everyday4 of 560 a day on Pro$0.0003
Kimi K3.1 ProOur pick: hard problems (shared)5 of 5125 a month on Pro$0.0050
On Pro: Pro's count on each model. Cost: what OpenRouter charged us per reply, on average.

Every limit is published. Paid plans also have a monthly fair-use limit on AI cost: Plus $3.75, 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.

Kimi K3.1 Pro next to each model on writing

Kimi K3.1 Pro against each model on the same 5 writing prompts. One run each, through OpenRouter: a wait depends on the provider and the load that day, so a lead under 10% counts as close.

  • Against Claude Fable 5.2: Kimi K3.1 Pro answered 2.0× later (6.9 s to 3.5 s), cost 2.3× less, and passed 5 of 5 to its 5.

  • Against Claude Fable 5.1: Kimi K3.1 Pro answered 1.7× later (6.9 s to 4.1 s), cost 2.7× less, and passed 5 of 5 to its 4. Only Kimi K3.1 Pro passed Argue both sides of free buses.

  • Against Claude Opus 5.5: Kimi K3.1 Pro answered 1.4× later (6.9 s to 4.8 s), cost 3.2× less, and passed 5 of 5 to its 5.

  • Against Claude Sonnet 5.5: Kimi K3.1 Pro answered 2.6× later (6.9 s to 2.7 s), cost 1.9× more, and passed 5 of 5 to its 4. Only Kimi K3.1 Pro passed Argue both sides of free buses.

  • Against Claude Sonnet 5: Kimi K3.1 Pro answered 1.9× later (6.9 s to 3.6 s), cost 1.3× more, and passed 5 of 5 to its 3. Only Kimi K3.1 Pro passed Announce a second bakery shop on LinkedIn; A product announcement with five rules.

  • Against Claude Haiku 5.5: Kimi K3.1 Pro answered 4.1× later (6.9 s to 1.7 s), cost 37.6× more, and passed 5 of 5 to its 3. Only Kimi K3.1 Pro passed Rewrite corporate jargon in plain words; Argue both sides of free buses.

  • Against Claude Haiku 4.5: Kimi K3.1 Pro answered 4.1× later (6.9 s to 1.7 s), cost 3.8× more, and passed 5 of 5 to its 4. Only Kimi K3.1 Pro passed A product announcement with five rules.

  • Against GPT-6.2 Astra: Kimi K3.1 Pro answered 1.4× later (6.9 s to 4.9 s), cost 1.8× less, and passed 5 of 5 to its 5.

  • Against GPT-6 Astra: Kimi K3.1 Pro answered 1.5× later (6.9 s to 4.7 s), cost 1.2× less, and passed 5 of 5 to its 4. Only Kimi K3.1 Pro passed Rewrite corporate jargon in plain words.

  • Against GPT-6.2 Sol: Kimi K3.1 Pro answered 3.8× later (6.9 s to 1.8 s), cost 2.2× more, and passed 5 of 5 to its 5.

  • Against GPT-6.1 Sol: Kimi K3.1 Pro answered 1.3× later (6.9 s to 5.5 s), cost 2.6× more, and passed 5 of 5 to its 5.

  • Against GPT-6 Sol: Kimi K3.1 Pro answered 2.1× later (6.9 s to 3.3 s), cost 3.8× more, and passed 5 of 5 to its 5.

  • Against GPT-6.1 Terra: Kimi K3.1 Pro answered 1.8× later (6.9 s to 3.9 s), cost 7.6× more, and passed 5 of 5 to its 5.

  • Against GPT-6 Luna: Kimi K3.1 Pro answered 2.7× later (6.9 s to 2.6 s), cost 89.0× more, and passed 5 of 5 to its 3. Only Kimi K3.1 Pro passed Rewrite corporate jargon in plain words; Argue both sides of free buses.

  • Against Gemini 4 Carbon: Kimi K3.1 Pro answered 3.0× sooner (6.9 s to 21.0 s), cost 1.3× less, and passed 5 of 5 to its 5.

  • Against Gemini 4 Argon: Kimi K3.1 Pro answered 1.1× sooner (6.9 s to 7.9 s), cost 3.6× more, and passed 5 of 5 to its 4. Only Kimi K3.1 Pro passed Argue both sides of free buses.

  • Against Gemini 4 Hydrogen: Kimi K3.1 Pro answered 1.1× sooner (6.9 s to 7.9 s), cost 1.3× less, and passed 5 of 5 to its 5.

  • Against Gemini 3.1 Pro (preview): Kimi K3.1 Pro answered within 10% of its time (6.9 s to 7.4 s), cost within 10% of it, and passed 5 of 5 to its 3. Only Kimi K3.1 Pro passed A product announcement with five rules; Argue both sides of free buses.

  • Against Gemini 3.8 Flash: Kimi K3.1 Pro answered 2.2× later (6.9 s to 3.1 s), cost 10.2× more, and passed 5 of 5 to its 5.

  • Against DeepSeek V4.1 Flash: Kimi K3.1 Pro answered 4.4× later (6.9 s to 1.6 s), cost 15.2× more, and passed 5 of 5 to its 4. Only Kimi K3.1 Pro passed Rewrite corporate jargon in plain words.

  • Against DeepSeek V4 Pro: Kimi K3.1 Pro answered 2.1× later (6.9 s to 3.3 s), cost 3.9× more, and passed 5 of 5 to its 5.

  • Against Grok 4.7: Kimi K3.1 Pro answered 1.2× sooner (6.9 s to 8.6 s), cost 1.1× less, and passed 5 of 5 to its 5.

  • Against Kimi K3: Kimi K3.1 Pro answered 1.3× later (6.9 s to 5.1 s), cost 1.5× more, and passed 5 of 5 to its 4. Only Kimi K3.1 Pro passed Argue both sides of free buses.

  • Against GLM 5.3: Kimi K3.1 Pro answered 4.5× later (6.9 s to 1.5 s), cost 5.8× more, and passed 5 of 5 to its 4. Only Kimi K3.1 Pro passed Argue both sides of free buses.

  • Against GLM 5.3 Flash: Kimi K3.1 Pro answered 6.8× later (6.9 s to 1.0 s), cost 30.5× more, and passed 5 of 5 to its 4. Only Kimi K3.1 Pro passed A product announcement with five rules.

  • Against Mistral Large 4: Kimi K3.1 Pro answered within 10% of its time (6.9 s to 7.2 s), cost 3.4× more, and passed 5 of 5 to its 5.

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.

How the runs were done, and every writing prompt.

Kimi K3.1 Pro, and writing, elsewhere

Questions

Is Kimi K3.1 Pro good for writing?

In our test runs it passed 5 of 5 writing prompts, tied 1st of the 27 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 Kimi K3.1 Pro use?

1 message each on Pro, where it has 125 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, GLM, and Mistral, in one chat.

See what a message costs before you send it. Every plan is month to month: cancel in two clicks. Sign in with an email link, no password.