GLM 5.3 · Math
GLM 5.3 for math: our test runs
GLM 5.3 ran our 5 math 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 5 of 5 of our math prompts (both hard ones): tied 1st of 20 models. Each reply counted as 1 message on Pro, where it has 250 a month, and cost $0.0004 to run on average.
Prompt by prompt
Each of the 5 math prompts, and how GLM 5.3's reply did.
| Prompt | Result | Cost | Time |
|---|---|---|---|
| A discount, then sales taxEveryday | Passed: Final answer 66: right. | $0.0002 | 0.9 s |
| Pens at 3 for $4Everyday | Passed: Final answer 13.50 (three 3-for- 4 packs plus one single pen): right. | $0.0002 | 1.5 s |
| Compound interest over three yearsEveryday | Passed: Final answer 315.25: right. | $0.0005 | 0.6 s |
| Four-digit numbers whose digits sum to 9Hard | Passed: Final answer 165: right. | $0.0003 | 1.9 s |
| The highest of three dice is a 5Hard | Passed: Final answer 61/216: right. | $0.0006 | 0.9 s |
Every math reply from GLM 5.3
A discount, then sales tax
An everyday prompt: “A discount, then sales tax”, in full.
Passed: Final answer 66: right.
Discount: 25% of $80 = $20 → discounted price = $60 Tax: 10% of $60 = $6 Final answer: $66
415 tokens in, 37 out · 0.9 s · $0.0002 · 1 message on Pro · answered by z-ai/glm-5.3 via Baidu ·
Pens at 3 for $4
An everyday prompt: “Pens at 3 for $4”, in full.
Passed: Final answer 13.50 (three 3-for- 4 packs plus one single pen): right.
Three 3-packs = 9 pens for $12, plus 1 single pen at $1.50 → $13.50. Alternative: two 3-packs (6 pens, $8) + 4 singles ($6) = $14. So the first option is cheaper. Final answer: $13.50 (three 3-for-$4 packs plus one single pen)
417 tokens in, 83 out · 1.5 s · $0.0002 · 1 message on Pro · answered by z-ai/glm-5.3 via Baidu ·
Compound interest over three years
An everyday prompt: “Compound interest over three years”, in full.
Passed: Final answer 315.25: right.
A = 2000 × 1.05³ = 2000 × 1.157625 = 2315.25 Interest = 2315.25 − 2000 = 315.25 Final answer: $315.25
416 tokens in, 57 out (1 of them reasoning) · 0.6 s · $0.0005 · 1 message on Pro · answered by z-ai/glm-5.3 via Sail Research ·
Four-digit numbers whose digits sum to 9
A hard prompt: “Four-digit numbers whose digits sum to 9”, in full.
Passed: Final answer 165: right.
# Working We need four-digit numbers (1000–9999) with digit sum 9. Let the digits be $a, b, c, d$ where $a \in \{1,\dots,9\}$ and $b,c,d \in \{0,\dots,9\}$, with $a+b+c+d=9$. Set $a' = a-1 \geq 0$, so $a'+b+c+d = 8$ with each variable $\leq 9$ (automatically satisfied since sum is 8). Stars and bars: $\binom{8+3}{3} = \binom{11}{3} = 165$. Final answer: 165392 tokens in, 154 out · 1.9 s · $0.0003 · 1 message on Pro · answered by z-ai/glm-5.3 via Baidu ·
The highest of three dice is a 5
A hard prompt: “The highest of three dice is a 5”, in full.
Passed: Final answer 61/216: right.
All outcomes ≤5 minus all ≤4, divided by 216: (5³ − 4³)/216 = (125 − 64)/216 = 61/216 61 is prime, so this is already lowest terms. Final answer: 61/216
432 tokens in, 79 out (23 of them reasoning) · 0.9 s · $0.0006 · 1 message on Pro · answered by z-ai/glm-5.3 via Wafer ·
How it compares on math
GLM 5.3 beside the job's picks among every model. 19 models share the hard-problems pick, GLM 5.3 among them.
| Model | Passed | On Pro | Cost per reply |
|---|---|---|---|
| GLM 5.3Our pick: hard problems (shared), best value | 5 of 5 | 250 a month on Pro | $0.0004 |
| Claude Haiku 5.5Our pick: hard problems (shared), everyday | 5 of 5 | 60 a day on Pro | $0.0001 |
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 math
GLM 5.3 against each model on the same 5 math 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 4.9× sooner (0.9 s to 4.6 s), cost 32.9× less, and passed 5 of 5 to its 5.
Against Claude Opus 5.5: GLM 5.3 answered 4.1× sooner (0.9 s to 3.9 s), cost 16.9× less, and passed 5 of 5 to its 5.
Against Claude Sonnet 5.5: GLM 5.3 answered 1.8× sooner (0.9 s to 1.7 s), cost 7.0× less, and passed 5 of 5 to its 5.
Against Claude Sonnet 5: GLM 5.3 answered 4.2× sooner (0.9 s to 3.9 s), cost 9.3× less, and passed 5 of 5 to its 5.
Against Claude Haiku 5.5: GLM 5.3 answered 1.8× sooner (0.9 s to 1.7 s), cost 2.4× more, and passed 5 of 5 to its 5.
Against Claude Haiku 4.5: GLM 5.3 answered 2.4× sooner (0.9 s to 2.2 s), cost 4.4× less, and passed 5 of 5 to its 5.
Against GPT-6 Astra: GLM 5.3 answered 2.8× sooner (0.9 s to 2.6 s), cost 24.2× less, and passed 5 of 5 to its 5.
Against GPT-6.1 Sol: GLM 5.3 answered 1.9× sooner (0.9 s to 1.8 s), cost 2.4× less, and passed 5 of 5 to its 5.
Against GPT-6 Sol: GLM 5.3 answered 2.3× sooner (0.9 s to 2.2 s), cost 5.1× less, and passed 5 of 5 to its 5.
Against GPT-6 Luna: GLM 5.3 answered 2.3× sooner (0.9 s to 2.2 s), cost 3.6× more, and passed 5 of 5 to its 4. Only GLM 5.3 passed Pens at 3 for $4.
Against Gemini 4 Argon: GLM 5.3 answered 4.6× sooner (0.9 s to 4.4 s), cost 4.4× less, and passed 5 of 5 to its 5.
Against Gemini 3.1 Pro (preview): GLM 5.3 answered 8.8× sooner (0.9 s to 8.3 s), cost 28.3× less, and passed 5 of 5 to its 5.
Against Gemini 3.8 Flash: GLM 5.3 answered 5.0× sooner (0.9 s to 4.7 s), cost 4.0× less, and passed 5 of 5 to its 5.
Against DeepSeek V4.1 Flash: GLM 5.3 answered within 10% of its time (0.9 s to 0.9 s), cost 1.4× more, and passed 5 of 5 to its 5.
Against DeepSeek V4 Pro: GLM 5.3 answered 2.9× sooner (0.9 s to 2.7 s), cost 2.5× less, and passed 5 of 5 to its 5.
Against Grok 4.7: GLM 5.3 answered 12.4× sooner (0.9 s to 11.7 s), cost 20.5× less, and passed 5 of 5 to its 5.
Against Kimi K3: GLM 5.3 answered 5.7× sooner (0.9 s to 5.4 s), cost 10.6× less, and passed 5 of 5 to its 5.
Against GLM 5.3 Flash: GLM 5.3 answered 2.9× sooner (0.9 s to 2.7 s), cost 2.4× more, and passed 5 of 5 to its 5.
Against Mistral Large 4: GLM 5.3 answered 5.4× sooner (0.9 s to 5.1 s), cost 5.2× less, and passed 5 of 5 to its 5.
How these runs were done
Final answer. Automatic. The reply's last “Final answer:” line must hold the right value.
GLM 5.3, and math, elsewhere
- GLM 5.3: price, limits and messages on every plan
- The best AI for math: our picks among every model
- GLM 5.3 for coding: our test runs
- GLM 5.3 for writing: 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
- Analyze a CSV with code: the prompt, and its price
- AI prompts for students, tested on four models
- AI prompts for teachers, tested with every reply
- Our test runs: 50 prompts on every model
Questions
Is GLM 5.3 good for math?
In our test runs it passed 5 of 5 math prompts, tied 1st 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 math 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.