Which is cheaper, Kimi K3 or Kimi K3.1 Pro?

Kimi K3.1 Pro is cheaper: $2 per million tokens read and $10 per million written, against $3 and $15 on Kimi K3. On a mix of 3 tokens read for each 1 written, Kimi K3 costs 1.5 times as much.

Which is better, Kimi K3 or Kimi K3.1 Pro?

Kimi K3 scores 157 on the Epoch Capabilities Index, an independent score from Epoch AI that combines over 50 benchmarks (GPT-5 is 150). Kimi K3.1 Pro has no independent score yet, so the only fair test is to run the same 10 tasks on both.

What does a support bot cost a month on Kimi K3 and Kimi K3.1 Pro?

At 2,000 replies a day, each reading 4,000 tokens (50 percent of them from cache) and writing 300, it costs $666 a month on Kimi K3 and $444 on Kimi K3.1 Pro: Kimi K3 costs 1.5 times as much for this job.

What does an AI agent task cost on Kimi K3 and Kimi K3.1 Pro?

One 40-turn task that reads 2,800,000 tokens in all, 90 percent of them from cache, and writes 40,000, costs $2.20 on Kimi K3 and $1.46 on Kimi K3.1 Pro. At 10 tasks a day that is $659 and $439 a month.

Which has the larger context window, Kimi K3 or Kimi K3.1 Pro?

Neither: both hold 1.05M tokens at once, about 1,572.9 pages of English text.

Does the same text take the same number of tokens on Kimi K3 and Kimi K3.1 Pro?

Yes, both are counted the same way: our lesson "How AI writes", 1,700 words long, is 2,055 tokens on each.

Is it cheaper to move from Kimi K3 to Kimi K3.1 Pro?

Yes, per token: Kimi K3.1 Pro costs $2 per million tokens read and $10 written, against $3 and $15 on Kimi K3, which costs 1.5 times as much for the same tokens. Run your own tasks on it before you switch, since answers, length and thinking can differ.

Can I pay less on Kimi K3 and Kimi K3.1 Pro for work that can wait?

Not with a batch price: neither maker's price page lists one for these models. A cache still lowers the price of text the model reads again.

Kimi K3 vs Kimi K3.1 Pro

Kimi K3.1 Pro is one of Moonshot AI's newest models and Kimi K3 an earlier one, so this is also what moving up costs or saves. Kimi K3.1 Pro is the cheaper of the 2: for the same tokens, 3 read for each 1 written, Kimi K3 costs 1.5 times as much. Below are both price lists, what 8 everyday jobs cost a month on each, and how many tokens the same text takes on each, which can change the answer.

Price your own job Compare 2 other models

For the same tokens: Kimi K3.1 Pro is cheaper; Kimi K3 costs 1.5 times as much
Cheaper to read: Kimi K3.1 Pro, $2 against $3
Cheaper to write: Kimi K3.1 Pro, $10 against $15
Scores higher: Not known, Kimi K3.1 Pro has no independent score yet
Larger window: The same, 1.05M tokens each

Price per million tokens, side by side

In US dollars, as each maker prints them. The better figure in each row is green.

	Kimi K3Moonshot AI · earlier model	Kimi K3.1 ProMoonshot AI
InputTokens the model reads	$3	$2
Cached inputRead again from the cache	$0.30	$0.20
Cache writePutting text in the cache	$3	$2
OutputTokens it writes, thinking included	$15	$10
Blend 3:13 tokens read for each 1 written	$6	$4
Batch inputAnswers collected later	Not listed	Not listed
Batch output	Not listed	Not listed
Context windowTokens it holds at once	1.05M	1.05M
Capability scoreEpoch Capabilities Index; GPT-5 is 150	157	Not listed
Longest answerTokens it writes in 1 reply	Not listed	Not listed

What 8 everyday jobs cost a month on each

Each job at its usual number of runs a day, for 30 days, at the standard price. The cheaper of the 2 is green; open any job in the calculator to change it.

Job	Kimi K3	Kimi K3.1 Pro	Difference
A chat question300 in, 500 out, 100 a day	$25.200.84 cents a run	$16.800.56 cents a run	1.5 times
A hard question, with thinking300 in, 1,400 out, 100 a day	$65.702.2 cents a run	$43.801.5 cents a run	1.5 times
A long report, summarised60,000 in, 2,000 out, 20 a day	$12621 cents a run	$84.0014 cents a run	1.5 times
A 40-turn agent task2,800,000 in, 40,000 out, 10 a day	$659$2.20 a run	$439$1.46 a run	1.5 times
A support bot's reply4,000 in, 300 out, 2,000 a day	$6661.1 cents a run	$4440.74 cents a run	1.5 times
Sort a message into a category250 in, 5 out, 50,000 a day	$1,2380.083 cents a run	$8250.055 cents a run	1.5 times
Read an invoice into fields1,500 in, 250 out, 1,000 a day	$2480.83 cents a run	$1650.55 cents a run	1.5 times
Review a pull request25,000 in, 4,500 out, 50 a day	$21414 cents a run	$1439.5 cents a run	1.5 times

Your own job on both

Start from a job, then set its size, how often it runs and whether the answers can wait. Both bills follow every change.

Start from a job

A short question and a few paragraphs back.

Read in 1 run

tokens

about 230 words

Read again from cache

Instructions or documents the model has read before cost far less the second time.

Written in 1 run, thinking included

tokens

about 380 words

Runs a day

A month of this job

Kimi K3$25.20
0.84 cents a run
Kimi K3.1 Pro$16.80
0.56 cents a run

Kimi K3.1 Pro saves $8.40 a month; Kimi K3 costs 1.5 times as much.

This job on every model

Why the price per token is only the start

Every maker prints 2 main prices: one for the tokens a model reads (input) and one for the tokens it writes (output). On most current models writing costs 3 to 5 times as much as reading, so which model is cheaper can depend on the job. A job that reads a long document and writes a line pays mostly the input price; a chat that writes paragraphs pays mostly the output price. That is why each comparison prices 8 jobs and not one figure.

The same text is not the same number of tokens on every model. Each maker cuts text its own way, and Anthropic says its models from Claude 4.7 on produce about 30 percent more tokens for the same text than earlier ones. A model with the lower price per token can still send the larger bill, so each comparison counts one of our lessons the way both makers count it.

Discounts change the order too. Work that can wait costs half on a batch at the makers that offer one, and text read again from a cache costs a quarter of the input price or less on every current model here. The price table on each comparison shows who offers what.

Thinking is billed as output, and models differ in how much they think and how long they answer. A price list cannot show that. Before you switch, run the same 10 tasks on both models and compare the bills.

Questions

Which is cheaper, Kimi K3 or Kimi K3.1 Pro?: Kimi K3.1 Pro is cheaper: $2 per million tokens read and $10 per million written, against $3 and $15 on Kimi K3. On a mix of 3 tokens read for each 1 written, Kimi K3 costs 1.5 times as much.
Which is better, Kimi K3 or Kimi K3.1 Pro?: Kimi K3 scores 157 on the Epoch Capabilities Index, an independent score from Epoch AI that combines over 50 benchmarks (GPT-5 is 150). Kimi K3.1 Pro has no independent score yet, so the only fair test is to run the same 10 tasks on both.
What does a support bot cost a month on Kimi K3 and Kimi K3.1 Pro?: At 2,000 replies a day, each reading 4,000 tokens (50 percent of them from cache) and writing 300, it costs $666 a month on Kimi K3 and $444 on Kimi K3.1 Pro: Kimi K3 costs 1.5 times as much for this job.
What does an AI agent task cost on Kimi K3 and Kimi K3.1 Pro?: One 40-turn task that reads 2,800,000 tokens in all, 90 percent of them from cache, and writes 40,000, costs $2.20 on Kimi K3 and $1.46 on Kimi K3.1 Pro. At 10 tasks a day that is $659 and $439 a month.
Which has the larger context window, Kimi K3 or Kimi K3.1 Pro?: Neither: both hold 1.05M tokens at once, about 1,572.9 pages of English text.
Does the same text take the same number of tokens on Kimi K3 and Kimi K3.1 Pro?: Yes, both are counted the same way: our lesson "How AI writes", 1,700 words long, is 2,055 tokens on each.
Is it cheaper to move from Kimi K3 to Kimi K3.1 Pro?: Yes, per token: Kimi K3.1 Pro costs $2 per million tokens read and $10 written, against $3 and $15 on Kimi K3, which costs 1.5 times as much for the same tokens. Run your own tasks on it before you switch, since answers, length and thinking can differ.
Can I pay less on Kimi K3 and Kimi K3.1 Pro for work that can wait?: Not with a batch price: neither maker's price page lists one for these models. A cache still lowers the price of text the model reads again.

Sources