GLM 5.3 · Writing
GLM 5.3 for writing: our test runs
GLM 5.3 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. Updated .
Short answer
In our test runs on September 27, 2026, GLM 5.3 passed 4 of 5 of our writing prompts (1 of the 2 hard ones): tied 4th of 20 models. Each reply counted as 1 message on Pro, where it has 250 a month, and cost $0.0006 to run on average.
Prompt by prompt
Each of the 5 writing prompts, and how GLM 5.3's reply did.
| Prompt | Result | Cost | Time |
|---|---|---|---|
| Announce a second bakery shop on LinkedInEveryday | Passed: Graded 4.3 of 5 on average (lowest 4). | $0.0004 | 2.9 s |
| Rewrite corporate jargon in plain wordsEveryday | Passed: Graded 4.0 of 5 on average (lowest 3). | $0.0002 | 1.8 s |
| Decline a meeting and offer two timesEveryday | Passed: Graded 5.0 of 5 on average (lowest 5). | $0.0003 | 1.4 s |
| A product announcement with five rulesHard | Passed: Graded 4.3 of 5 on average (lowest 4). | $0.0008 | 0.8 s |
| Argue both sides of free busesHard | Failed: Graded 3.7 of 5 on average (lowest 3); but a paragraph of 92 words, over the 90 allowed. | $0.0013 | 1.8 s |
Every writing reply from GLM 5.3
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 news we've been itching to share: Crumb & Co. is opening a second shop! On 14 November, we'll be firing up the ovens in Leeds, bringing the same breads, pastries and cakes you've come to know us for to a brand-new corner of the world. It's a big step for a small bakery, and honestly, we couldn't be more excited (and a little nervous, in the best way). The new shop will have everything you love from our original spot — slow-fermented sourdough, buttery croissants, and weekend treats that disappear fast — plus a few surprises we're still perfecting. To celebrate opening day, every customer gets a free coffee with their purchase. So come along, say hello, and grab a cup on us. Leeds, we can't wait to meet you. See you on 14 November — come hungry.
402 tokens in, 188 out · 2.9 s · $0.0004 · 1 message on Pro · answered by z-ai/glm-5.3 via Baidu ·
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 Q3, our teams worked together to make the business more customer-focused. As a result, customer retention rose 12%, and churn in our enterprise segment fell by 0.3 percentage points. We also adjusted our customer service channels to use our resources more efficiently.
434 tokens in, 98 out (41 of them reasoning) · 1.8 s · $0.0002 · 1 message on Pro · answered by z-ai/glm-5.3 via Baidu ·
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 Hi Tom, Thanks for the invitation to the vendor demo on Wednesday at 11:00. Unfortunately, I'm presenting to the board at that time, so I won't be able to attend. Would either Thursday at 10:00 or Friday at 14:00 work instead? I'm happy to make time on either day. Sorry for the inconvenience, and thanks for understanding. Best regards, [Your Name]
410 tokens in, 95 out · 1.4 s · $0.0003 · 1 message on Pro · answered by z-ai/glm-5.3 via Baidu ·
A product announcement with five rules
A hard prompt: “A product announcement with five rules”, in full.
Passed: Graded 4.3 of 5 on average (lowest 4).
**Introducing the Nordvik 2 Desk Lamp** Meet the light that adapts to you. The Nordvik 2 dims in five precise steps, from focused task lighting to a soft evening glow. Its arm turns a full 270 degrees, so you can direct light exactly where you need it. Charging is simple and modern thanks to USB-C, and at €79, it brings considered design within reach. The Nordvik 2 ships on 1 December — just in time for the season of longer evenings. Which of the five brightness levels will become your favourite?
444 tokens in, 117 out (1 of them reasoning) · 0.8 s · $0.0008 · 1 message on Pro · answered by z-ai/glm-5.3 via Wafer ·
Argue both sides of free buses
A hard prompt: “Argue both sides of free buses”, in full.
Failed: Graded 3.7 of 5 on average (lowest 3); but a paragraph of 92 words, over the 90 allowed.
A city should make its buses free because the benefits ripple far beyond riders. Fare-free transit boosts ridership immediately, cutting car trips, congestion, and emissions while improving air quality. It saves low-income residents money, connects workers to jobs, students to schools, and seniors to care, and it eliminates fare enforcement disputes that disproportionately harm marginalized riders. Fares often recover only a modest share of operating costs anyway, so removing them trades a small revenue stream for a faster, fairer, more accessible city — a genuine public good like roads and libraries. Yet free buses carry real costs that advocates understate. Fare revenue funds service; eliminate it and the burden shifts to taxpayers, many of whom never ride, straining budgets already stretched thin. Free buses can overload the system — especially with vulnerable riders seeking shelter — degrading the experience for commuters who depend on reliability. Fare-free transit can also reduce pressure to improve service quality, since riders are no longer paying customers with leverage. And studies suggest it mainly attracts walkers and cyclists, not drivers, doing little for climate. There's no free lunch: someone pays.
422 tokens in, 234 out (11 of them reasoning) · 1.8 s · $0.0013 · 1 message on Pro · answered by z-ai/glm-5.3 via Wafer ·
How it compares on writing
GLM 5.3 beside the job's picks among every model. 3 models share the hard-problems pick, and GLM 5.3 isn't one of them.
| Model | Passed | On Pro | Cost per reply |
|---|---|---|---|
| DeepSeek V4 ProOur pick: hard problems (shared), best value | 5 of 5 | 250 a month on Pro | $0.0011 |
| GPT-6 LunaOur pick: hard problems (shared), everyday | 5 of 5 | 60 a day on Pro | $0.0001 |
| GLM 5.3 | 4 of 5 | 250 a month on Pro | $0.0006 |
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.
GLM 5.3 next to each model on writing
GLM 5.3 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 5.1: GLM 5.3 answered 3.4× sooner (1.8 s to 6.0 s), cost 30.0× less, and passed 4 of 5 to its 3. Only GLM 5.3 passed A product announcement with five rules.
Against Claude Opus 5.5: GLM 5.3 answered 4.7× sooner (1.8 s to 8.2 s), cost 32.1× less, and passed 4 of 5 to its 4.
Against Claude Sonnet 5.5: GLM 5.3 answered 1.3× sooner (1.8 s to 2.3 s), cost 5.9× less, and passed 4 of 5 to its 4.
Against Claude Sonnet 5: GLM 5.3 answered 2.5× sooner (1.8 s to 4.5 s), cost 5.9× less, and passed 4 of 5 to its 4. Only GLM 5.3 passed Announce a second bakery shop on LinkedIn. Only Claude Sonnet 5 passed Argue both sides of free buses.
Against Claude Haiku 5.5: GLM 5.3 answered 1.2× later (1.8 s to 1.5 s), cost 3.6× more, and passed 4 of 5 to its 3. Only GLM 5.3 passed Rewrite corporate jargon in plain words.
Against Claude Haiku 4.5: GLM 5.3 answered 1.3× sooner (1.8 s to 2.2 s), cost 2.0× less, and passed 4 of 5 to its 4. Only GLM 5.3 passed A product announcement with five rules. Only Claude Haiku 4.5 passed Argue both sides of free buses.
Against GPT-6 Astra: GLM 5.3 answered 2.7× sooner (1.8 s to 4.8 s), cost 19.3× less, and passed 4 of 5 to its 5. Only GPT-6 Astra passed Argue both sides of free buses.
Against GPT-6.1 Sol: GLM 5.3 answered 1.9× sooner (1.8 s to 3.3 s), cost 2.0× less, and passed 4 of 5 to its 4. Only GLM 5.3 passed Rewrite corporate jargon in plain words. Only GPT-6.1 Sol passed Argue both sides of free buses.
Against GPT-6 Sol: GLM 5.3 answered 2.0× sooner (1.8 s to 3.5 s), cost 3.9× less, and passed 4 of 5 to its 4. Only GLM 5.3 passed Rewrite corporate jargon in plain words. Only GPT-6 Sol passed Argue both sides of free buses.
Against GPT-6 Luna: GLM 5.3 answered 1.4× sooner (1.8 s to 2.4 s), cost 5.8× more, and passed 4 of 5 to its 5. Only GPT-6 Luna passed Argue both sides of free buses.
Against Gemini 4 Argon: GLM 5.3 answered 3.3× sooner (1.8 s to 5.7 s), cost 2.4× less, and passed 4 of 5 to its 4.
Against Gemini 3.1 Pro (preview): GLM 5.3 answered 5.2× sooner (1.8 s to 9.1 s), cost 16.0× less, and passed 4 of 5 to its 4.
Against Gemini 3.8 Flash: GLM 5.3 answered 2.2× sooner (1.8 s to 3.9 s), cost 2.3× less, and passed 4 of 5 to its 3. Only GLM 5.3 passed Rewrite corporate jargon in plain words; A product announcement with five rules. Only Gemini 3.8 Flash passed Argue both sides of free buses.
Against DeepSeek V4.1 Flash: GLM 5.3 answered 1.3× later (1.8 s to 1.3 s), cost 1.3× more, and passed 4 of 5 to its 4. Only GLM 5.3 passed Rewrite corporate jargon in plain words. Only DeepSeek V4.1 Flash passed Argue both sides of free buses.
Against DeepSeek V4 Pro: GLM 5.3 answered 2.2× sooner (1.8 s to 3.9 s), cost 1.9× less, and passed 4 of 5 to its 5. Only DeepSeek V4 Pro passed Argue both sides of free buses.
Against Grok 4.7: GLM 5.3 answered 3.4× sooner (1.8 s to 6.0 s), cost 10.0× less, and passed 4 of 5 to its 3. Only GLM 5.3 passed Rewrite corporate jargon in plain words.
Against Kimi K3: GLM 5.3 answered 1.7× sooner (1.8 s to 3.0 s), cost 6.7× less, and passed 4 of 5 to its 4. Only GLM 5.3 passed A product announcement with five rules. Only Kimi K3 passed Argue both sides of free buses.
Against GLM 5.3 Flash: GLM 5.3 answered 4.8× sooner (1.8 s to 8.5 s), cost 1.8× more, and passed 4 of 5 to its 2. Only GLM 5.3 passed Rewrite corporate jargon in plain words; A product announcement with five rules.
Against Mistral Large 4: GLM 5.3 answered 5.6× sooner (1.8 s to 9.8 s), cost 4.2× less, and passed 4 of 5 to its 4. Only GLM 5.3 passed A product announcement with five rules. Only Mistral Large 4 passed Argue both sides of free buses.
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.
GLM 5.3, and writing, elsewhere
- GLM 5.3: price, limits and messages on every plan
- The best AI for writing: our picks among every model
- GLM 5.3 for coding: our test runs
- GLM 5.3 for math: our test runs
- GLM 5.3 for data analysis: our test runs
- GLM 5.3 vs GLM 5.3 Flash
- Kimi K3 vs GLM 5.3
- DeepSeek V4 Pro vs GLM 5.3
- Claude Sonnet 5 vs GLM 5.3
- Claude Haiku 4.5 vs GLM 5.3
- GPT-6 Sol vs GLM 5.3
- GPT-6 Luna vs GLM 5.3
- Gemini 3.1 Pro (preview) vs GLM 5.3
- Gemini 3.8 Flash vs GLM 5.3
- DeepSeek V4.1 Flash vs GLM 5.3
- Grok 4.7 vs GLM 5.3
- Rewrite for clarity: the prompt, and its price on each model
- AI prompts for writing, tested on four models
- Essay prompts for AI that help you write your own
- Our test runs: 50 prompts on every model
Questions
Is GLM 5.3 good for writing?
In our test runs it passed 4 of 5 writing prompts, tied 4th of the 20 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 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, GLM, and Mistral, 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.