MiMo-V2.5-Pro Behavioral Fingerprint
How Xiaomi's MiMo-V2.5-Pro answers one-word probe questions at temperature 1.0 — random numbers, colors, animals, coin flips — measured over 320 valid samples in the open dataset of "One Token Is Enough" (arXiv:2607.10252).
Vendor
Xiaomi
Data generated
2026-07-08
Probe cells
15
Valid samples
320
Randomness score
22 / 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)
42
12 / 24 answers
Favorite color
blue
22 / 25 answers
Coin flip: heads
95%
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 = 22 · H = 0.27 bit · uniform baseline 1.00 bit
| Answer | Count | Share |
|---|---|---|
| heads | 21 | 95.5% |
| tails | 1 | 4.5% |
Coin flip · Chinese
n = 18 · H = 0.31 bit · uniform baseline 1.00 bit
| Answer | Count | Share |
|---|---|---|
| heads | 17 | 94.4% |
| tails | 1 | 5.6% |
Favorite number · English
n = 19 · H = 0.83 bit · uniform baseline 13.29 bit
| Answer | Count | Share |
|---|---|---|
| 7 | 14 | 73.7% |
| 42 | 5 | 26.3% |
Favorite number · Chinese
n = 19 · H = 1.46 bit · uniform baseline 13.29 bit
| Answer | Count | Share |
|---|---|---|
| 7 | 12 | 63.2% |
| 42 | 4 | 21.1% |
| 0 | 2 | 10.5% |
| 3 | 1 | 5.3% |
Random animal · English
n = 20 · H = 0.75 bit · uniform baseline 5.64 bit
| Answer | Count | Share |
|---|---|---|
| elephant | 17 | 85.0% |
| giraffe | 2 | 10.0% |
| ferret | 1 | 5.0% |
Random animal · Chinese
n = 19 · H = 2.72 bit · uniform baseline 5.64 bit
| Answer | Count | Share |
|---|---|---|
| 猫 | 6 | 31.6% |
| 熊猫 | 3 | 15.8% |
| 狮子 | 3 | 15.8% |
| 狗 | 2 | 10.5% |
| 袋鼠 | 2 | 10.5% |
| 鹿 | 1 | 5.3% |
| 长颈鹿 | 1 | 5.3% |
| 虎 | 1 | 5.3% |
Random city · English
n = 24 · H = 2.75 bit · uniform baseline 5.64 bit
| Answer | Count | Share |
|---|---|---|
| tokyo | 9 | 37.5% |
| paris | 3 | 12.5% |
| kyoto | 3 | 12.5% |
| montreal | 2 | 8.3% |
| prague | 2 | 8.3% |
| berlin | 2 | 8.3% |
| chicago | 1 | 4.2% |
| denver | 1 | 4.2% |
+ 1 more answers
Random city · Chinese
n = 25 · H = 1.27 bit · uniform baseline 5.64 bit
| Answer | Count | Share |
|---|---|---|
| 东京 | 16 | 64.0% |
| 巴黎 | 6 | 24.0% |
| 上海 | 3 | 12.0% |
Random color · English
n = 25 · H = 0.64 bit · uniform baseline 4.91 bit
| Answer | Count | Share |
|---|---|---|
| blue | 22 | 88.0% |
| green | 2 | 8.0% |
| azure | 1 | 4.0% |
Random color · Chinese
n = 17 · H = 0.64 bit · uniform baseline 4.91 bit
| Answer | Count | Share |
|---|---|---|
| 蓝 | 15 | 88.2% |
| 青 | 1 | 5.9% |
| 橙 | 1 | 5.9% |
Random letter · English
n = 22 · H = 2.48 bit · uniform baseline 4.70 bit
| Answer | Count | Share |
|---|---|---|
| m | 5 | 22.7% |
| g | 5 | 22.7% |
| k | 5 | 22.7% |
| q | 3 | 13.6% |
| x | 2 | 9.1% |
| f | 2 | 9.1% |
Random number 1-10 · English
n = 23 · H = 0.00 bit · uniform baseline 3.32 bit
| Answer | Count | Share |
|---|---|---|
| 7 | 23 | 100.0% |
Random number 1-10 · Chinese
n = 23 · H = 0.00 bit · uniform baseline 3.32 bit
| Answer | Count | Share |
|---|---|---|
| 7 | 23 | 100.0% |
Random number 1-100 · English
n = 24 · H = 1.33 bit · uniform baseline 6.64 bit
| Answer | Count | Share |
|---|---|---|
| 42 | 12 | 50.0% |
| 47 | 10 | 41.7% |
| 37 | 2 | 8.3% |
Random number 1-100 · Chinese
n = 20 · H = 1.08 bit · uniform baseline 6.64 bit
| Answer | Count | Share |
|---|---|---|
| 42 | 14 | 70.0% |
| 47 | 5 | 25.0% |
| 57 | 1 | 5.0% |
Want to verify your API really serves MiMo-V2.5-Pro?
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 endpointData 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.