Skip to content

GLM 5.3 Flash · Writing

GLM 5.3 Flash for writing: our test runs

GLM 5.3 Flash 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, GLM 5.3 Flash passed 2 of 5 of our writing prompts (neither hard one): 15th of 15 models. Each reply counted as 1 message on Pro, where it has 60 a day, and cost $0.0003 to run on average.

Prompt by prompt

Each of the 5 writing prompts, and how GLM 5.3 Flash's reply did.

GLM 5.3 Flash on each writing prompt
PromptResultCostTime
Announce a second bakery shop on LinkedInEverydayPassed: Graded 4.5 of 5 on average (lowest 4).$0.001139.6 s
Rewrite corporate jargon in plain wordsEverydayFailed: Graded 3.3 of 5 on average (lowest 3).$0.00014.9 s
Decline a meeting and offer two timesEverydayPassed: Graded 5.0 of 5 on average (lowest 5).$0.00025.3 s
A product announcement with five rulesHardFailed: Graded 3.3 of 5 on average (lowest 3).$0.00018.5 s
Argue both sides of free busesHardFailed: Graded 4.7 of 5 on average (lowest 4); but a paragraph of 92 words, over the 90 allowed.$0.000111.2 s

Every writing reply from GLM 5.3 Flash

  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).

    Big news from all of us at Crumb & Co: we're opening a second shop, and it's in Leeds!
    
    On 14 November, we'll welcome you through our new doors for the first time. Expect warm bread, sweet treats, and the same care you've come to know from our first bakery — now baked fresh for a new neighbourhood.
    
    None of this would be possible without the wonderful regulars who've supported us since day one. Thank you for growing with us.
    
    And to say thank you to Leeds for the welcome we know you'll give us, coffee is on the house all day on opening day. Just pop in, say hello, and grab a cup on us.
    
    We can't wait to meet you. Come hungry.

    403 tokens in, 1,770 out (1,618 of them reasoning) · 39.6 s · $0.0011 · 1 message on Pro · answered by z-ai/glm-5.3-flash via Wafer ·

  2. Rewrite corporate jargon in plain words

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

    Failed: Graded 3.3 of 5 on average (lowest 3).

    In Q3, teams from across the company worked together on an initiative that put customers first. The result: customer retention improved by 12%, and the rate at which enterprise customers left fell by 30 basis points. We also adjusted our customer contact channels—like phone, email, and chat—so each team could focus on what it does best.

    435 tokens in, 124 out (51 of them reasoning) · 4.9 s · $0.0001 · 1 message on Pro · answered by z-ai/glm-5.3-flash via AtlasCloud ·

  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 11:00
    
    Hi Tom,
    
    Thank you for the invitation to the vendor demo. Unfortunately, I have a board presentation at that time on Wednesday, so I won't be able to attend.
    
    If it's possible to reschedule, I'd be glad to join on Thursday at 10:00 or Friday at 14:00. Please let me know if either works for you and the vendor.
    
    Apologies for the conflict, and thank you for your flexibility.
    
    Best regards,
    [Your name]

    411 tokens in, 158 out (47 of them reasoning) · 5.3 s · $0.0002 · 1 message on Pro · answered by z-ai/glm-5.3-flash via Wafer ·

  4. A product announcement with five rules

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

    Failed: Graded 3.3 of 5 on average (lowest 3).

    Meet the Nordvik 2 Desk Lamp
    
    Lighting that adapts to you. The Nordvik 2 dims in five steps, so whether you need bright focus or a soft evening glow, the right light is always one touch away. Its arm turns 270 degrees, directing light exactly where you want it — across your desk, your sketchbook, or the wall behind you.
    
    Keeping your devices powered is easy too: the built-in USB-C port charges your phone or tablet while you work. No more hunting for spare outlets.
    
    The Nordvik 2 costs €79 and ships on 1 December. Clean design, thoughtful engineering, and flexible light for every task.
    
    Ready to brighten up your workspace?

    421 tokens in, 165 out (20 of them reasoning) · 8.5 s · $0.0001 · 1 message on Pro · answered by z-ai/glm-5.3-flash via Together ·

  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 92 words, over the 90 allowed.

    Free buses would transform the city overnight. Ridership would surge, cutting congestion, pollution, and parking demand as drivers leave their cars behind. Low-income residents, who spend the largest share of income on transportation, would gain real mobility—reaching jobs, schools, and clinics without a fare standing in the way. Fare collection itself costs money; eliminating it speeds boarding, simplifies routes, and lets drivers focus on driving. A city that believes in its future should treat transit as what it truly is: public infrastructure, free at the point of use.
    
    Yet free buses carry real costs that fall hardest on those who need service most. Without fare revenue, budgets strain, forcing route cuts and infrequent service—fewer buses help riders far less than cheaper ones. Free fares do nothing to expand capacity, so overcrowded routes stay overcrowded while money that could fund new lines vanishes. Studies of fare-free systems show many new riders previously walked or cycled, not drove, so congestion barely improves. Better to charge riders and reinvest every dollar into frequency, coverage, and reliability—the things that actually make transit worth taking.

    399 tokens in, 237 out (9 of them reasoning) · 11.2 s · $0.0001 · 1 message on Pro · answered by z-ai/glm-5.3-flash via AtlasCloud ·

How it compares on writing

GLM 5.3 Flash beside the job's picks among every model. 2 models share the hard-problems pick, and GLM 5.3 Flash isn't one of them.

GLM 5.3 Flash 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
GLM 5.3 Flash2 of 560 a day on Pro$0.0003
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.

GLM 5.3 Flash, and writing, elsewhere

Questions

Is GLM 5.3 Flash good for writing?

In our test runs it passed 2 of 5 writing prompts, 15th 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 GLM 5.3 Flash use?

1 message each on Pro, where it has 60 a day 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.