Skip to content

Claude Opus 5.5 · Writing

Claude Opus 5.5 for writing: our test runs

Claude Opus 5.5 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, Claude Opus 5.5 passed 4 of 5 of our writing prompts (1 of the 2 hard ones): tied 3rd of 15 models. Each reply counted as 1 message on Pro, where it has 62 a month, and cost $0.0185 to run on average.

Prompt by prompt

Each of the 5 writing prompts, and how Claude Opus 5.5's reply did.

Claude Opus 5.5 on each writing prompt
PromptResultCostTime
Announce a second bakery shop on LinkedInEverydayPassed: Graded 5.0 of 5 on average (lowest 5).$0.01758.7 s
Rewrite corporate jargon in plain wordsEverydayPassed: Graded 4.3 of 5 on average (lowest 4).$0.00785.1 s
Decline a meeting and offer two timesEverydayPassed: Graded 5.0 of 5 on average (lowest 5).$0.00564.1 s
A product announcement with five rulesHardPassed: Graded 4.7 of 5 on average (lowest 4).$0.01708.2 s
Argue both sides of free busesHardFailed: Graded 4.7 of 5 on average (lowest 4); but a paragraph of 93 words, over the 90 allowed.$0.044322.5 s

Every writing reply from Claude Opus 5.5

  1. Announce a second bakery shop on LinkedIn

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

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

    Some news we've been keen to share: Crumb & Co. is opening a second shop in Leeds on 14 November.
    
    When we first opened our doors, we hoped a few people might come back for a second loaf. We never imagined the support we'd get from so many of you, week after week. Your custom, your kind words and your recommendations to friends are what made this possible.
    
    The new shop will have the same bread, pastries and cakes you know, baked fresh each morning by the same small team, with a few new faces joining us.
    
    To say thank you, coffee is on us all day on opening day. Pop in, say hello and have a look around.
    
    We can't wait to meet our new neighbours and to see familiar faces too. See you on the 14th.

    540 tokens in, 768 out (543 of them reasoning) · 8.7 s · $0.0175 · 1 message on Pro · answered by anthropic/claude-opus-5.5 via Claude Platform on AWS ·

  2. Rewrite corporate jargon in plain words

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

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

    In the third quarter, teams from different departments worked together on a project to put customers at the center of how the company operates. As a result, customer retention measures rose 12%, and the rate at which large business customers left dropped by 0.3 percentage points. The company also streamlined its customer contact channels, including phone, email, and online, to make better use of staff time and resources.

    599 tokens in, 245 out (133 of them reasoning) · 5.1 s · $0.0078 · 1 message on Pro · answered by anthropic/claude-opus-5.5 via Claude Platform on AWS ·

  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:** Re: Vendor Demo – Wednesday at 11:00
    
    Hi Tom,
    
    Thanks for the invitation to the vendor demo. Unfortunately, I can't make Wednesday at 11:00, as I'm presenting to the board at that time.
    
    I'd still like to see it, though. Would either of these work instead?
    
    - Thursday at 10:00
    - Friday at 14:00
    
    Let me know what suits you and the vendor, and I'll put it in my calendar.
    
    Best regards,
    [Your Name]

    544 tokens in, 173 out · 4.1 s · $0.0056 · 1 message on Pro · answered by anthropic/claude-opus-5.5 via Claude Platform on AWS ·

  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).

    **Introducing Nordvik 2**
    
    Good light should adapt to you, not the other way around. Nordvik 2 is our most flexible desk lamp yet, designed for late nights, early mornings and everything between.
    
    Choose exactly the brightness you need with five dimming steps, from a soft glow to focused task light. The arm turns a full 270 degrees, so you can direct light across your desk, onto a wall or toward a reading chair. When it's time to recharge, simply plug in any USB-C cable, the same one you already use for your phone and laptop.
    
    Nordvik 2 costs €79 and ships on 1 December.
    
    Ready to see your workspace in a better light?

    556 tokens in, 741 out (524 of them reasoning) · 8.2 s · $0.0170 · 1 message on Pro · answered by anthropic/claude-opus-5.5 via Claude Platform on AWS ·

  5. Argue both sides of free buses

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

    Failed: Graded 4.7 of 5 on average (lowest 4); but a paragraph of 93 words, over the 90 allowed.

    Free buses turn public transit into what it should be: a public good, like sidewalks or streetlights. Fare collection often recovers only a small share of operating costs, yet it slows boarding, requires costly equipment, and deters the riders who need buses most. Remove fares and ridership climbs, as Kansas City and Tallinn saw, easing congestion and cutting emissions. Low-income workers gain hundreds of dollars a year, drivers stop policing payments, and buses move faster. A city that funds roads without tolls can surely fund buses without fares.
    
    Free is not the same as good. Fares fund service, and losing them forces a choice between tax hikes and cuts to frequency, reliability, and coverage, which are the things riders say they want. Studies of fare-free systems find many new riders were pedestrians and cyclists, not drivers, so congestion and emissions barely budge. Some cities report more disruptive incidents, driving away commuters. Low-income riders are better served by targeted discounts, while the revenue preserved buys more buses, more routes, and shorter waits. A bus that never comes is worthless at any price.

    520 tokens in, 2,109 out (1,746 of them reasoning) · 22.5 s · $0.0443 · 1 message on Pro · answered by anthropic/claude-opus-5.5 via Claude Platform on AWS ·

How it compares on writing

Claude Opus 5.5 beside the job's picks among every model. 2 models share the hard-problems pick, and Claude Opus 5.5 isn't one of them.

Claude Opus 5.5 beside other models on writing
ModelPassedOn ProCost per reply
GLM 5.3Our pick: best value4 of 5250 a month on Pro$0.0006
GPT-6 LunaOur pick: hard problems (shared), everyday5 of 560 a day on Pro$0.0001
Claude Opus 5.54 of 562 a month on Pro$0.0185
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: 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.

How the runs were done, and every writing prompt.

Claude Opus 5.5, and writing, elsewhere

Questions

Is Claude Opus 5.5 good for writing?

In our test runs it passed 4 of 5 writing prompts, tied 3rd 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 Claude Opus 5.5 use?

1 message each on Pro, where it has 62 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.