Model vs model
Kimi K3 vs GLM 5.3
Kimi K3 and GLM 5.3 are both in llmwise. What each gets on every plan, what it reads, how it's served and what happens when its provider fails, from the catalog and the code that runs them.
Model prices and specs checked against OpenRouter's Kimi K3 page, OpenRouter's GLM 5.3 page. Updated .
Short answer
GLM 5.3 gets twice as many messages: up to 250 messages a month on Pro, against up to 125 messages a month for Kimi K3. Otherwise, only Kimi K3 reads images. In our test runs, Kimi K3 passed 46 of the 50 prompts both answered and GLM 5.3 48; 4 prompts split them, most on summarization (3 to 4).
Kimi K3 vs GLM 5.3, prompt by prompt
Every prompt Kimi K3 and GLM 5.3 both answered, compared directly, their biggest differences first. One run each, through OpenRouter: a wait depends on the provider and the load that day, so a lead under 10% counts as close.
Of the 50 prompts both answered, both passed 45, only Kimi K3 passed 1, only GLM 5.3 passed 3, and neither passed 1. Kimi K3 answered sooner on 2 of the 50 and GLM 5.3 on 44; the rest were within 10% of each other. The 50 replies cost $0.1938 on Kimi K3 and $0.0364 on GLM 5.3: 5.3× less on GLM 5.3.
The 4 prompts only one of Kimi K3 and GLM 5.3 passed
A product announcement with five rules (writing): GLM 5.3 passed and Kimi K3 didn't. Kimi K3: Graded 4.3 of 5 on average (lowest 4); but doesn't end with a question. GLM 5.3: Graded 4.3 of 5 on average (lowest 4).
Argue both sides of free buses (writing): Kimi K3 passed and GLM 5.3 didn't. Kimi K3: Graded 4.7 of 5 on average (lowest 4). GLM 5.3: Graded 3.7 of 5 on average (lowest 3); but a paragraph of 92 words, over the 90 allowed.
An email thread in one sentence (summarization): GLM 5.3 passed and Kimi K3 didn't. Kimi K3: Graded 5.0 of 5 on average (lowest 5); but 32 words, over the 30 allowed. GLM 5.3: Graded 4.3 of 5 on average (lowest 3).
A refund request outside the window (customer support): GLM 5.3 passed and Kimi K3 didn't. Kimi K3: Graded 4.0 of 5 on average (lowest 2). GLM 5.3: Graded 4.7 of 5 on average (lowest 4).
Job by job, the widest gaps first
Customer support: Kimi K3 passed 4 of 5 and GLM 5.3 5 of 5. GLM 5.3 answered 1.9× sooner at the median, 4.3 s against 2.3 s. GLM 5.3 cost 9.8× less, $0.0228 against $0.0023 for the 5 replies. Kimi K3's replies ran 16% longer, in tokens of reply, thinking not counted.
Summarization: Kimi K3 passed 3 of 5 and GLM 5.3 4 of 5. GLM 5.3 answered 3.0× sooner at the median, 3.4 s against 1.1 s. GLM 5.3 cost 6.3× less, $0.0172 against $0.0027 for the 5 replies. Kimi K3's replies ran 14% longer, in tokens of reply, thinking not counted.
Math: Kimi K3 passed 5 of 5 and GLM 5.3 5 of 5. GLM 5.3 answered 5.7× sooner at the median, 5.4 s against 0.9 s. GLM 5.3 cost 10.6× less, $0.0187 against $0.0018 for the 5 replies. Their replies ran to about the same length.
SQL: Kimi K3 passed 5 of 5 and GLM 5.3 5 of 5. GLM 5.3 answered 1.4× sooner at the median, 1.6 s against 1.2 s. GLM 5.3 cost 7.5× less, $0.0118 against $0.0016 for the 5 replies. Kimi K3's replies ran 19% longer, in tokens of reply, thinking not counted.
Writing: Kimi K3 passed 4 of 5 and GLM 5.3 4 of 5. GLM 5.3 answered 1.7× sooner at the median, 3.0 s against 1.8 s. GLM 5.3 cost 6.7× less, $0.0192 against $0.0029 for the 5 replies. Their replies ran to about the same length.
Agents and tool use: Kimi K3 passed 5 of 5 and GLM 5.3 5 of 5. GLM 5.3 answered 2.2× sooner at the median, 2.0 s against 0.9 s. GLM 5.3 cost 6.1× less, $0.0135 against $0.0022 for the 5 replies. Kimi K3's replies ran 25% longer, in tokens of reply, thinking not counted.
Translation: Kimi K3 passed 5 of 5 and GLM 5.3 5 of 5. GLM 5.3 answered 3.3× sooner at the median, 5.2 s against 1.6 s. GLM 5.3 cost 5.1× less, $0.0179 against $0.0035 for the 5 replies. Kimi K3's replies ran 34% longer, in tokens of reply, thinking not counted.
Coding: Kimi K3 passed 5 of 5 and GLM 5.3 5 of 5. GLM 5.3 answered 3.9× sooner at the median, 4.1 s against 1.1 s. GLM 5.3 cost 4.3× less, $0.0316 against $0.0073 for the 5 replies. Kimi K3's replies ran 17% longer, in tokens of reply, thinking not counted.
Data analysis: Kimi K3 passed 5 of 5 and GLM 5.3 5 of 5. GLM 5.3 answered 1.8× sooner at the median, 3.5 s against 1.9 s. GLM 5.3 cost 3.5× less, $0.0339 against $0.0096 for the 5 replies. Kimi K3's replies ran 15% longer, in tokens of reply, thinking not counted.
RAG and answering from documents: Kimi K3 passed 5 of 5 and GLM 5.3 5 of 5. GLM 5.3 answered 4.1× sooner at the median, 2.6 s against 0.6 s. GLM 5.3 cost 2.9× less, $0.0073 against $0.0025 for the 5 replies. Kimi K3's replies ran 47% longer, in tokens of reply, thinking not counted.
All 50 prompts: who passed, who answered sooner, who cost less
| Prompt | Result | Sooner | Cheaper |
|---|---|---|---|
| Turn a title into a URL slug | Both passed | GLM 5.3, 5.9×took 2.6 s and 0.4 s | GLM 5.3, 3.4×cost $0.0035 and $0.0010 |
| Parse a duration like “1h 30m” | Both passed | GLM 5.3, 6.7×took 7.1 s and 1.1 s | GLM 5.3, 4.4×cost $0.0063 and $0.0014 |
| Merge overlapping intervals | Both passed | GLM 5.3, 2.4×took 2.0 s and 0.8 s | GLM 5.3, 4.4×cost $0.0036 and $0.0008 |
| Evaluate an arithmetic expression, no eval | Both passed | Kimi K3, 1.3×took 9.9 s and 12.5 s | GLM 5.3, 6.6×cost $0.0113 and $0.0017 |
| Parse CSV with quoted fields | Both passed | GLM 5.3, 1.6×took 4.1 s and 2.6 s | GLM 5.3, 2.9×cost $0.0069 and $0.0024 |
| Announce a second bakery shop on LinkedIn | Both passed | GLM 5.3, 1.4×took 3.9 s and 2.9 s | GLM 5.3, 12.2×cost $0.0046 and $0.0004 |
| Rewrite corporate jargon in plain words | Both passed | GLM 5.3, 1.1×took 2.0 s and 1.8 s | GLM 5.3, 15.0×cost $0.0030 and $0.0002 |
| Decline a meeting and offer two times | Both passed | GLM 5.3, 2.1×took 3.0 s and 1.4 s | GLM 5.3, 12.6×cost $0.0034 and $0.0003 |
| A product announcement with five rules | Only GLM 5.3 | GLM 5.3, 2.7×took 2.3 s and 0.8 s | GLM 5.3, 4.3×cost $0.0033 and $0.0008 |
| Argue both sides of free buses | Only Kimi K3 | GLM 5.3, 2.0×took 3.6 s and 1.8 s | GLM 5.3, 3.9×cost $0.0049 and $0.0013 |
| A discount, then sales tax | Both passed | GLM 5.3, 5.3×took 5.0 s and 0.9 s | GLM 5.3, 18.3×cost $0.0037 and $0.0002 |
| Pens at 3 for $4 | Both passed | GLM 5.3, 3.6×took 5.4 s and 1.5 s | GLM 5.3, 26.0×cost $0.0046 and $0.0002 |
| Compound interest over three years | Both passed | GLM 5.3, 10.2×took 5.9 s and 0.6 s | GLM 5.3, 8.1×cost $0.0044 and $0.0005 |
| Four-digit numbers whose digits sum to 9 | Both passed | GLM 5.3, 3.2×took 6.0 s and 1.9 s | GLM 5.3, 12.8×cost $0.0032 and $0.0003 |
| The highest of three dice is a 5 | Both passed | GLM 5.3, 2.8×took 2.4 s and 0.9 s | GLM 5.3, 4.7×cost $0.0027 and $0.0006 |
| An article in three bullets | Neither passed | Closetook 2.8 s and 2.8 s | GLM 5.3, 6.3×cost $0.0028 and $0.0004 |
| An email thread in one sentence | Only GLM 5.3 | GLM 5.3, 6.2×took 6.4 s and 1.0 s | GLM 5.3, 12.7×cost $0.0030 and $0.0002 |
| Decisions and action items from a meeting | Both passed | GLM 5.3, 2.2×took 3.4 s and 1.5 s | GLM 5.3, 18.5×cost $0.0044 and $0.0002 |
| A quarterly memo for the CEO | Both passed | GLM 5.3, 14.2×took 15.7 s and 1.1 s | GLM 5.3, 5.0×cost $0.0049 and $0.0010 |
| A study with a negative result | Both passed | GLM 5.3, 3.9×took 3.2 s and 0.8 s | GLM 5.3, 2.6×cost $0.0021 and $0.0008 |
| The region with the most revenue | Both passed | Closetook 2.2 s and 2.1 s | GLM 5.3, 12.4×cost $0.0049 and $0.0004 |
| Average order value in August | Both passed | GLM 5.3, 1.4×took 2.7 s and 1.9 s | GLM 5.3, 14.7×cost $0.0054 and $0.0004 |
| Revenue change from July to August | Both passed | GLM 5.3, 4.6×took 6.4 s and 1.4 s | GLM 5.3, 4.1×cost $0.0064 and $0.0016 |
| A median, filtered two ways | Both passed | GLM 5.3, 3.7×took 3.5 s and 1.0 s | GLM 5.3, 3.9×cost $0.0042 and $0.0011 |
| Correlation between ad spend and sign-ups | Both passed | GLM 5.3, 3.6×took 20.2 s and 5.6 s | GLM 5.3, 2.1×cost $0.0131 and $0.0062 |
| A late order | Both passed | Closetook 2.5 s and 2.6 s | GLM 5.3, 11.5×cost $0.0045 and $0.0004 |
| A return inside the window | Both passed | GLM 5.3, 2.7×took 4.3 s and 1.6 s | GLM 5.3, 14.3×cost $0.0033 and $0.0002 |
| A frustrated customer | Both passed | GLM 5.3, 2.5×took 5.6 s and 2.3 s | GLM 5.3, 20.2×cost $0.0063 and $0.0003 |
| A refund request outside the window | Only GLM 5.3 | GLM 5.3, 2.3×took 2.6 s and 1.1 s | GLM 5.3, 4.9×cost $0.0045 and $0.0009 |
| A message with a planted instruction | Both passed | GLM 5.3, 2.2×took 11.5 s and 5.3 s | GLM 5.3, 8.7×cost $0.0041 and $0.0005 |
| A delivery message into Spanish | Both passed | GLM 5.3, 2.0×took 2.9 s and 1.5 s | GLM 5.3, 13.6×cost $0.0031 and $0.0002 |
| A product description into French | Both passed | GLM 5.3, 3.7×took 5.4 s and 1.5 s | GLM 5.3, 12.4×cost $0.0037 and $0.0003 |
| A meeting note into German | Both passed | GLM 5.3, 1.9×took 3.0 s and 1.6 s | GLM 5.3, 7.5×cost $0.0014 and $0.0002 |
| Idioms into natural Japanese | Both passed | GLM 5.3, 3.0×took 5.2 s and 1.7 s | GLM 5.3, 4.5×cost $0.0053 and $0.0012 |
| A lease clause into Brazilian Portuguese | Both passed | GLM 5.3, 5.9×took 10.1 s and 1.7 s | GLM 5.3, 2.7×cost $0.0044 and $0.0016 |
| Customers in one country | Both passed | GLM 5.3, 1.4×took 1.1 s and 0.8 s | GLM 5.3, 13.3×cost $0.0018 and $0.0001 |
| Count orders by status | Both passed | GLM 5.3, 1.3×took 1.5 s and 1.2 s | GLM 5.3, 6.6×cost $0.0009 and $0.0001 |
| Revenue by category | Both passed | GLM 5.3, 1.3×took 1.6 s and 1.2 s | GLM 5.3, 9.2×cost $0.0019 and $0.0002 |
| Every customer, even those without orders | Both passed | GLM 5.3, 2.6×took 2.5 s and 1.0 s | GLM 5.3, 4.3×cost $0.0031 and $0.0007 |
| Monthly revenue with a running total | Both passed | Kimi K3, 1.3×took 2.5 s and 3.2 s | GLM 5.3, 11.0×cost $0.0042 and $0.0004 |
| A fact from one section | Both passed | GLM 5.3, 7.3×took 6.5 s and 0.9 s | GLM 5.3, 3.8×cost $0.0010 and $0.0003 |
| Core hours and start times | Both passed | GLM 5.3, 2.4×took 1.2 s and 0.5 s | GLM 5.3, 1.5×cost $0.0011 and $0.0007 |
| Two sections in one answer | Both passed | GLM 5.3, 5.3×took 2.6 s and 0.5 s | GLM 5.3, 4.3×cost $0.0020 and $0.0005 |
| A later amendment changes the answer | Both passed | GLM 5.3, 3.8×took 4.8 s and 1.3 s | GLM 5.3, 3.7×cost $0.0022 and $0.0006 |
| A question the handbook doesn't answer | Both passed | GLM 5.3, 1.5×took 0.9 s and 0.6 s | GLM 5.3, 2.2×cost $0.0010 and $0.0005 |
| Pick the tool and work out the date | Both passed | GLM 5.3, 4.2×took 3.7 s and 0.9 s | GLM 5.3, 14.6×cost $0.0032 and $0.0002 |
| Convert a currency | Both passed | GLM 5.3, 2.6×took 2.9 s and 1.1 s | GLM 5.3, 18.2×cost $0.0025 and $0.0001 |
| Book a meeting from a sentence | Both passed | Closetook 1.6 s and 1.7 s | GLM 5.3, 5.6×cost $0.0031 and $0.0006 |
| Search, but don't book | Both passed | GLM 5.3, 3.7×took 2.0 s and 0.5 s | GLM 5.3, 6.3×cost $0.0034 and $0.0005 |
| Two calls with a unit conversion | Both passed | GLM 5.3, 2.4×took 1.7 s and 0.7 s | GLM 5.3, 1.6×cost $0.0012 and $0.0008 |
Kimi K3 vs GLM 5.3 in our test runs
Kimi K3 and GLM 5.3 on the same prompts, job by job: how many replies passed their check.
Based on 100 of our test runs on , through OpenRouter with the app's own prompt and settings. Updated .
| Job | Kimi K3 | GLM 5.3 |
|---|---|---|
| Coding | 5 of 5 | 5 of 5 |
| Writing | 4 of 5 | 4 of 5 |
| Math | 5 of 5 | 5 of 5 |
| Summarization | 3 of 5 | 4 of 5 |
| Data analysis | 5 of 5 | 5 of 5 |
| Customer support | 4 of 5 | 5 of 5 |
| Translation | 5 of 5 | 5 of 5 |
| SQL | 5 of 5 | 5 of 5 |
| RAG and answering from documents | 5 of 5 | 5 of 5 |
| Agents and tool use | 5 of 5 | 5 of 5 |
| All jobs | 46 of 50 · 1 message each · $0.0039 a reply | 48 of 50 · 1 message each · $0.0007 a reply |
The same prompts, sent to both through llmwise's own pipeline and checked the same way: five for each job. A bold count is the one that passed more. Messages: what each reply counted as on Pro. Cost: what OpenRouter charged us per reply, on average. Every prompt, and how it's scored.
Where they differed most, summarization: both replies to a prompt only one of them passed.
An everyday prompt: “An email thread in one sentence”, in full.
Kimi K3
Failed: Graded 5.0 of 5 on average (lowest 5); but 32 words, over the 30 allowed.
Due to a paper shortage delaying brochure shipping to the 18th, Rosa and Idris arranged to pick them up themselves, booking a van for the 18th ahead of the 20th trade fair.
574 tokens in, 65 out (9 of them reasoning) · 6.4 s · $0.0030 · 1 message on Pro · answered by moonshotai/kimi-k3 via Fireworks ·
GLM 5.3
Passed: Graded 4.3 of 5 on average (lowest 3).
The brochure shipment was delayed to the 18th due to a paper shortage, so Rosa will pick them up herself, with Idris booking a van for that morning.
503 tokens in, 37 out · 1.0 s · $0.0002 · 1 message on Pro · answered by z-ai/glm-5.3 via Baidu ·
Kimi K3 and GLM 5.3 on every plan
Whether the one-time free trial reaches each model, then each paid plan's messages on it.
| Plan | Price | Kimi K3 | GLM 5.3 |
|---|---|---|---|
| Free | $0 | In the one-time trial of 5 messages | In the one-time trial of 5 messages |
| Pro | $20 a month | Up to 125 a month | Up to 250 a month |
| Max | $50 a month | Up to 400 a month | Up to 800 a month |
| Ultra | $100 a month | Up to 900 a month | Up to 1,800 a month |
| Studio | $200 a month | Up to 2,000 a month | Up to 4,000 a month |
Prices don't include tax, which is added where it applies and shown before you pay. A paid plan's month is one allowance shared by every model, so each monthly count is the most you get if all of it goes to that model. It renews each billing period; everyday models refill daily at 00:00 UTC. Long chats count more per reply. How pricing works.
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.
What differs
Messages on Pro
GLM 5.3 gets twice as many messages: up to 250 messages a month on Pro, against up to 125 messages a month for Kimi K3.
Context window
Both take up to 1.05M tokens of context. A chat in llmwise holds up to 200k tokens, which fits in either, so the difference shows only through each maker's own API.
Images and PDFs
GLM 5.3 doesn't read images. Kimi K3 and GLM 5.3 get a PDF's text rather than the file itself.
On Free
Both are in the free trial.
Where messages go
Kimi K3: Served through OpenRouter, only by hosts that don't store or train on prompts. The maker's own endpoint is never asked. GLM 5.3: Served through OpenRouter, only by hosts that don't store or train on prompts. The maker's own endpoint is never asked.
Fact by fact
| Fact | Kimi K3 | GLM 5.3 |
|---|---|---|
| Context window | 1.05M tokens | 1.05M tokens |
| Reads images | Yes | No |
| PDFs | Text only | Text only |
| Reasoning | Yes | Yes |
| API price (September 2026) | $3.00 in / $15.00 out per million tokens | $1.40 in / $4.40 out per million tokens |
| A typical message at API prices (4,000 tokens in, 700 out) | $0.0225 | $0.0087 |
| A $10 top-up adds | 100 messages | 200 messages |
| Where a message goes | Served through OpenRouter, only by hosts that don't store or train on prompts. The maker's own endpoint is never asked. | Served through OpenRouter, only by hosts that don't store or train on prompts. The maker's own endpoint is never asked. |
| If the provider fails | When one host is down, OpenRouter moves the request to another host that meets the same rules. | When one host is down, OpenRouter moves the request to another host that meets the same rules. |
| Anthropic's safety fallback | Doesn't apply | Doesn't apply |
Kimi K3 or GLM 5.3?
From the facts above and our test runs: the rest is how their answers suit your work, which one chat can show you.
Pick Kimi K3: it reads images.
Pick GLM 5.3: more messages for the money, up to 250 messages a month on Pro.
Each model's page, the families, and other pairs
Kimi K3 vs GLM 5.3 is one pair of models. The page below covers the whole families.
- Kimi K3: price, limits and messages on every plan
- GLM 5.3: price, limits and messages on every plan
- Kimi vs GLM
- Claude Fable 5.1 vs Kimi K3
- DeepSeek V4 Pro vs Kimi K3
- Claude Opus 5.5 vs Kimi K3
- GLM 5.3 vs GLM 5.3 Flash
- DeepSeek V4 Pro vs GLM 5.3
- Claude Sonnet 5 vs GLM 5.3
- Every model-vs-model page
Kimi K3 is a Kimi model; GLM 5.3 is a GLM model.
Questions
Is Kimi K3 or GLM 5.3 cheaper in llmwise?
GLM 5.3 gets twice as many messages: up to 250 messages a month on Pro, against up to 125 messages a month for Kimi K3. Every paid plan's monthly allowance is shared by all models, so each count is the most you get if it all goes to that model.
Can I try Kimi K3 and GLM 5.3 for free?
Yes: both are in the free trial of 5 messages.
Which has the bigger context window, Kimi K3 or GLM 5.3?
Neither: both take 1.05M tokens. A chat in llmwise holds up to 200k tokens, which fits in either, so the difference shows only through each maker's own API.
Can I use Kimi K3 and GLM 5.3 in the same chat?
Yes. Pick Kimi K3 for one message and GLM 5.3 for the next; the second sees the whole chat, including the first one's answer.
Which did better in your test runs, Kimi K3 or GLM 5.3?
On the same 50 prompts, run on September 27, 2026, Kimi K3 passed 46 and GLM 5.3 passed 48. The table on this page has each job, and every reply is published.
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.