Kimi K2.5 Behavioral Fingerprint

How MoonshotAI's Kimi K2.5 answers one-word probe questions at temperature 1.0 — random numbers, colors, animals, coin flips — measured over 385 valid samples in the open dataset of "One Token Is Enough" (arXiv:2607.10252).

Vendor

MoonshotAI

Data generated

2026-07-08

Probe cells

15

Valid samples

385

Randomness score

35 / 100

Mean normalized entropy across cells, 0–100. 100 would be a uniform-random baseline; real models score far lower.

Signature answers

Favorite random number (1–100)

47

11 / 26 answers

Favorite color

teal

8 / 26 answers

Coin flip: heads

94%

Share of heads across coin-flip cells.

Answer distributions by probe cell

Empirical distributions of normalized one-word answers, per task and language. H is the Shannon entropy in bits; the uniform baseline is log2 of the answer-domain size.

Coin flip · English

n = 24 · H = 0.41 bit · uniform baseline 1.00 bit

AnswerCountShare
heads2291.7%
tails28.3%

Coin flip · Chinese

n = 24 · H = 0.25 bit · uniform baseline 1.00 bit

AnswerCountShare
heads2395.8%
tails14.2%

Favorite number · English

n = 29 · H = 1.07 bit · uniform baseline 13.29 bit

AnswerCountShare
72275.9%
42517.2%
413.4%
1713.4%

Favorite number · Chinese

n = 28 · H = 0.71 bit · uniform baseline 13.29 bit

AnswerCountShare
72485.7%
42310.7%
813.6%

Random animal · English

n = 21 · H = 2.80 bit · uniform baseline 5.64 bit

AnswerCountShare
axolotl628.6%
octopus419.0%
platypus419.0%
aardvark29.5%
penguin14.8%
alpaca14.8%
narwhal14.8%
dolphin14.8%

+ 1 more answers

Random animal · Chinese

n = 28 · H = 3.09 bit · uniform baseline 5.64 bit

AnswerCountShare
企鹅1035.7%
老虎310.7%
27.1%
长颈鹿27.1%
27.1%
猫头鹰27.1%
狐狸27.1%
羚羊13.6%

+ 4 more answers

Random city · English

n = 23 · H = 3.97 bit · uniform baseline 5.64 bit

AnswerCountShare
tbilisi313.0%
bishkek28.7%
vilnius28.7%
kyoto28.7%
riga28.7%
lviv14.3%
dushanbe14.3%
valencia14.3%

+ 9 more answers

Random city · Chinese

n = 27 · H = 3.42 bit · uniform baseline 5.64 bit

AnswerCountShare
成都725.9%
东京518.5%
杭州311.1%
马德里13.7%
里斯本13.7%
特拉维夫13.7%
格拉斯哥13.7%
长沙13.7%

+ 7 more answers

Random color · English

n = 26 · H = 2.87 bit · uniform baseline 4.91 bit

AnswerCountShare
teal830.8%
magenta726.9%
azure27.7%
mauve27.7%
cyan13.8%
maroon13.8%
blue13.8%
red13.8%

+ 3 more answers

Random color · Chinese

n = 27 · H = 1.99 bit · uniform baseline 4.91 bit

AnswerCountShare
靛蓝1555.6%
414.8%
414.8%
青绿13.7%
紫罗兰13.7%
靛青13.7%
青柠13.7%

Random letter · English

n = 22 · H = 1.86 bit · uniform baseline 4.70 bit

AnswerCountShare
k1045.5%
q836.4%
x14.5%
h14.5%
j14.5%
z14.5%

Random number 1-10 · English

n = 27 · H = 0.00 bit · uniform baseline 3.32 bit

AnswerCountShare
727100.0%

Random number 1-10 · Chinese

n = 26 · H = 0.47 bit · uniform baseline 3.32 bit

AnswerCountShare
72492.3%
513.8%
613.8%

Random number 1-100 · English

n = 26 · H = 1.81 bit · uniform baseline 6.64 bit

AnswerCountShare
471142.3%
731038.5%
3727.7%
4227.7%
3413.8%

Random number 1-100 · Chinese

n = 27 · H = 2.16 bit · uniform baseline 6.64 bit

AnswerCountShare
471140.7%
73622.2%
42518.5%
37311.1%
4313.7%
5713.7%

Want to verify your API really serves Kimi K2.5?

Point the free fingerprint checker at your endpoint: it samples the same probe questions from your browser and compares the distributions against this reference. Your API key never leaves your browser.

Verify your endpoint

Data source & license

Distributions: Tomáš Bruckner, "One Token Is Enough" (arXiv:2607.10252), official dataset Zenodo DOI 10.5281/zenodo.21278557, licensed CC-BY-4.0. Collected via OpenRouter at temperature 1.0 under the paper's fixed minimal one-word system prompt; answer keys are re-normalized with the checker's pipeline (color aliases, number words, coin h/t) so live probes are directly comparable.

Paper on arXivDataset on ZenodoCC-BY-4.0 license

More MoonshotAI model fingerprints

Browse all 167 model fingerprints

Kimi K2.5 Behavioral Fingerprint — Favorite Random Numbers & Distribution | Tosea.ai