Skip to content

Claude Haiku 4.5 · Writing

Claude Haiku 4.5 for writing: our test runs

Claude Haiku 4.5 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 Haiku 4.5 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 250 a month, and cost $0.0011 to run on average.

Prompt by prompt

Each of the 5 writing prompts, and how Claude Haiku 4.5's reply did.

Claude Haiku 4.5 on each writing prompt
PromptResultCostTime
Announce a second bakery shop on LinkedInEverydayPassed: Graded 4.0 of 5 on average (lowest 3).$0.00132.6 s
Rewrite corporate jargon in plain wordsEverydayPassed: Graded 4.3 of 5 on average (lowest 4).$0.00081.6 s
Decline a meeting and offer two timesEverydayPassed: Graded 5.0 of 5 on average (lowest 5).$0.00102.0 s
A product announcement with five rulesHardFailed: Graded 3.7 of 5 on average (lowest 3).$0.00122.2 s
Argue both sides of free busesHardPassed: Graded 4.0 of 5 on average (lowest 4).$0.00143.4 s

Every writing reply from Claude Haiku 4.5

  1. Announce a second bakery shop on LinkedIn

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

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

    We're thrilled to announce that Crumb & Co. is expanding! Our second shop opens in Leeds on 14 November.
    
    Since we started, you've supported us with every sourdough, pastry, and cup of coffee. This new location means more of you can enjoy fresh-baked goodness closer to home.
    
    To say thank you, we're offering complimentary coffee all day on opening day. Pop by, say hello, and see what we've built.
    
    Our Leeds team is ready to welcome you with the same warmth and quality that defines Crumb & Co. Same recipes, same care, same commitment to baking something special each morning.
    
    Here's to new beginnings—and more reasons to visit us.
    
    See you on 14 November.

    420 tokens in, 169 out · 2.6 s · $0.0013 · 1 message on Pro · answered by anthropic/claude-haiku-4.5 via Amazon Bedrock ·

  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).

    Here's a clearer version:
    
    Our Q3 initiative brought different departments together to focus on customer needs. The results were strong: we kept 12% more customers and reduced customer loss by 0.3 percentage points among large businesses. We also streamlined our customer contact channels to work more efficiently.
    
    (77 words)

    462 tokens in, 73 out · 1.6 s · $0.0008 · 1 message on Pro · answered by anthropic/claude-haiku-4.5 via Amazon Bedrock ·

  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 Invitation – Wednesday
    
    Hi Tom,
    
    Thanks for the invitation to the vendor demo on Wednesday at 11:00. Unfortunately, I have a board presentation scheduled at that time and won't be able to attend.
    
    I'm interested in seeing the demo though. Would either of these times work for you instead?
    
    - Thursday at 10:00
    - Friday at 14:00
    
    Let me know what suits your schedule best, and I'll get it on my calendar.
    
    Best regards

    425 tokens in, 112 out · 2.0 s · $0.0010 · 1 message on Pro · answered by anthropic/claude-haiku-4.5 via Amazon Bedrock ·

  4. 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 3).

    # Introducing the Nordvik 2 Desk Lamp
    
    Meet the Nordvik 2, designed for modern workspaces. This versatile lamp features a 270-degree rotating arm that positions light exactly where you need it. Five-step dimming lets you adjust brightness from soft ambient to focused task lighting.
    
    Charging is effortless with USB-C connectivity, keeping your desk cable-free and clutter-free. The minimalist design complements any interior while delivering reliable performance.
    
    At just €79, the Nordvik 2 offers exceptional value for professionals and students alike. Pre-order now for delivery on 1 December.
    
    Ready to upgrade your workspace with smarter lighting?

    437 tokens in, 154 out · 2.2 s · $0.0012 · 1 message on Pro · answered by anthropic/claude-haiku-4.5 via Amazon Bedrock ·

  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).

    Free buses reduce traffic congestion and emissions while improving public health through increased mobility. They eliminate collection costs and fare evasion losses, funding themselves partly through reduced road maintenance. Equity improves dramatically—low-income residents gain economic mobility and job access. Cities like Luxembourg prove feasibility. The modest expense is offset by economic productivity gains, reduced healthcare costs, and environmental benefits that create long-term savings.
    
    However, free transit strains already limited budgets without guaranteed ridership increases to offset costs. Maintenance deteriorates as systems become overcrowded without capacity expansion funding. Cities risk losing accountability—free services often receive less operational scrutiny. Revenue vanishes precisely when transit needs modernization most. Targeted subsidies for low-income riders prove more efficient than universal free fares. Better solutions include congestion pricing for private vehicles and progressive fare structures, reserving resources for quality service rather than unsustainable giveaways.

    414 tokens in, 195 out · 3.4 s · $0.0014 · 1 message on Pro · answered by anthropic/claude-haiku-4.5 via Amazon Bedrock ·

How it compares on writing

Claude Haiku 4.5 beside the job's picks among every model. 2 models share the hard-problems pick, and Claude Haiku 4.5 isn't one of them.

Claude Haiku 4.5 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 Haiku 4.54 of 5250 a month on Pro$0.0011
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 Haiku 4.5, and writing, elsewhere

Questions

Is Claude Haiku 4.5 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 Haiku 4.5 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.