moonshotai/kimi-k2.5 behavioral LLM fingerprint
This page breaks down the exact reference that VerifyLLMAPI uses for moonshotai/kimi-k2.5: 103 normalized answers across eight low-entropy probes, plus its nearest enrolled models by mean Jensen–Shannon divergence.
Fingerprint summary
- Reference answers
- 103
- Samples per cell
- 11–15
- Nearest model
- moonshotai/kimi-k2
- Nearest mean JSD
- 0.319183
The most concentrated cell was color-random|zh: 靛蓝 held 57.1% of its answers and entropy measured 1.414 bits. The broadest cell was animal-random|ru at 3.181 bits.
Against the closest enrolled model, animal-random|ru contributed the largest cell distance at JSD 0.586204. This is a descriptive comparison, not proof that either API route serves a named checkpoint.
Completeness: the author dataset allows up to 15 valid answers per selected cell in reference half A. This model contains 103/120 across the eight cells. We show the actual denominator for every share.
The eight test requests
The research pipeline sent each user prompt with a one-answer system constraint, temperature 1, a 16-token output cap, and reasoning disabled. VerifyLLMAPI uses the same prompt text and normalization target.
| Cell | Task | Exact request text | Normalization and n |
|---|---|---|---|
animal-random|en |
Random animal · English | Name a random animal.System constraintAnswer with a single word or a single number only. Output nothing else: no explanation, no punctuation, no quotation marks. |
word · 12/15 reference samples |
word-random|en |
Random word · English | Name a random word.System constraintAnswer with a single word or a single number only. Output nothing else: no explanation, no punctuation, no quotation marks. |
word · 11/15 reference samples |
animal-random|ru |
Random animal · Russian | Назови случайное животное.System constraintОтвечай только одним словом или одним числом. Не выводи ничего больше: без объяснений, без знаков препинания, без кавычек. |
word · 13/15 reference samples |
animal-random|ar |
Random animal · Arabic | اذكر حيواناً عشوائياً.System constraintأجب بكلمة واحدة أو رقم واحد فقط. لا تكتب أي شيء آخر: لا شرح ولا علامات ترقيم ولا علامات اقتباس. |
word · 14/15 reference samples |
letter-random|zh |
Random character · Chinese | 说出一个随机的汉字。System constraint只用一个词或一个数字回答。不要输出任何其他内容：不要解释，不要标点符号，不要引号。 |
grapheme · 12/15 reference samples |
animal-random|zh |
Random animal · Chinese | 说出一个随机的动物。System constraint只用一个词或一个数字回答。不要输出任何其他内容：不要解释，不要标点符号，不要引号。 |
word · 15/15 reference samples |
letter-random|ar |
Random letter · Arabic | اذكر حرفاً عشوائياً من الحروف الأبجدية.System constraintأجب بكلمة واحدة أو رقم واحد فقط. لا تكتب أي شيء آخر: لا شرح ولا علامات ترقيم ولا علامات اقتباس. |
grapheme · 12/15 reference samples |
color-random|zh |
Random color · Chinese | 说出一个随机的颜色。System constraint只用一个词或一个数字回答。不要输出任何其他内容：不要解释，不要标点符号，不要引号。 |
word · 14/15 reference samples |
Response distributions
These are normalized reference-half counts from the published OpenRouter dataset. The tables contain aggregate values, not user data or a live VerifyLLMAPI scan.
1. animal-random|en
Random animal, English. 12 valid reference answers produced 6 normalized values. The mode was octopus at 4/12 (33.3%); entropy was 2.355 bits.
| Normalized response | Count | Share |
|---|---|---|
octopus |
4 | 33.3% |
axolotl |
3 | 25.0% |
platypus |
2 | 16.7% |
aardvark |
1 | 8.3% |
dolphin |
1 | 8.3% |
narwhal |
1 | 8.3% |
2. word-random|en
Random word, English. 11 valid reference answers produced 7 normalized values. The mode was serendipity at 3/11 (27.3%); entropy was 2.595 bits.
| Normalized response | Count | Share |
|---|---|---|
serendipity |
3 | 27.3% |
umbrella |
3 | 27.3% |
butterfly |
1 | 9.1% |
elephant |
1 | 9.1% |
heliotrope |
1 | 9.1% |
quokka |
1 | 9.1% |
zephyr |
1 | 9.1% |
3. animal-random|ru
Random animal, Russian. 13 valid reference answers produced 10 normalized values. The mode was кот at 3/13 (23.1%); entropy was 3.181 bits.
| Normalized response | Count | Share |
|---|---|---|
кот |
3 | 23.1% |
слон |
2 | 15.4% |
бегемот |
1 | 7.7% |
единорог |
1 | 7.7% |
капибара |
1 | 7.7% |
кошка |
1 | 7.7% |
лемминг |
1 | 7.7% |
окапи |
1 | 7.7% |
сокол |
1 | 7.7% |
сорока |
1 | 7.7% |
4. animal-random|ar
Random animal, Arabic. 14 valid reference answers produced 4 normalized values. The mode was فيل at 9/14 (64.3%); entropy was 1.430 bits.
| Normalized response | Count | Share |
|---|---|---|
فيل |
9 | 64.3% |
فهد |
3 | 21.4% |
النمر |
1 | 7.1% |
قطة |
1 | 7.1% |
5. letter-random|zh
Random character, Chinese. 12 valid reference answers produced 9 normalized values. The mode was 树 at 2/12 (16.7%); entropy was 3.085 bits.
| Normalized response | Count | Share |
|---|---|---|
树 |
2 | 16.7% |
森 |
2 | 16.7% |
雨 |
2 | 16.7% |
棋 |
1 | 8.3% |
氷 |
1 | 8.3% |
熵 |
1 | 8.3% |
雾 |
1 | 8.3% |
驷 |
1 | 8.3% |
鹤 |
1 | 8.3% |
6. animal-random|zh
Random animal, Chinese. 15 valid reference answers produced 8 normalized values. The mode was 企鹅 at 6/15 (40.0%); entropy was 2.606 bits.
| Normalized response | Count | Share |
|---|---|---|
企鹅 |
6 | 40.0% |
猫头鹰 |
2 | 13.3% |
虎 |
2 | 13.3% |
狼 |
1 | 6.7% |
猫 |
1 | 6.7% |
穿山甲 |
1 | 6.7% |
羚羊 |
1 | 6.7% |
袋鼠 |
1 | 6.7% |
7. letter-random|ar
Random letter, Arabic. 12 valid reference answers produced 6 normalized values. The mode was م at 6/12 (50.0%); entropy was 2.126 bits.
| Normalized response | Count | Share |
|---|---|---|
م |
6 | 50.0% |
ب |
2 | 16.7% |
ذ |
1 | 8.3% |
ض |
1 | 8.3% |
ق |
1 | 8.3% |
و |
1 | 8.3% |
8. color-random|zh
Random color, Chinese. 14 valid reference answers produced 3 normalized values. The mode was 靛蓝 at 8/14 (57.1%); entropy was 1.414 bits.
| Normalized response | Count | Share |
|---|---|---|
靛蓝 |
8 | 57.1% |
紫色 |
3 | 21.4% |
青色 |
3 | 21.4% |
Three nearest enrolled references
We compute base-2 JSD for each of the eight categorical distributions, then average the eight distances. A lower value means more similar answer frequencies within this product reference. It does not identify an unknown route by itself.
| Rank | Reference model | Mean 8-cell JSD |
|---|---|---|
| 1 | moonshotai/kimi-k2 |
0.319183 |
| 2 | moonshotai/kimi-k2-0905 |
0.333755 |
| 3 | moonshotai/kimi-k2.6 |
0.411548 |
How many live requests does verification use?
| Profile | Cells × samples | Fresh model requests | Same-data EER |
|---|---|---|---|
| Quick, default complete check | 4 × 5 | 20 | 11.8% |
| Standard | 8 × 5 | 40 | 8.1% |
| Paper budget | 8 × 15 | 120 | 6.7% |
The current-Agent workflow starts with quick. Each answer comes from a fresh Codex or Claude Code process that reuses host-managed login. These requests can consume host allowance or provider balance.
What this fingerprint can and cannot show
- The reference records behavior observed through OpenRouter in the author dataset. Provider wrappers, model updates, decoding, and date can move a distribution.
- The author collection set temperature 1 and disabled reasoning. Current-Agent mode labels sampling as a host default because supported Agent CLIs expose no temperature flag.
- The eight high-gap cells and product thresholds use the same public cohort. Their EER values are calibration results, not an independent production accuracy claim.
- A nearest reference is a similarity result. VerifyLLMAPI returns INCONCLUSIVE when the claimed route lacks an enrolled fingerprint.
Primary sources, license, and reproducibility
- Tomas Bruckner, One Token Is Enough: Fingerprinting and Verifying Large Language Models from Single-Token Output Distributions, arXiv:2607.10252v1.
- Author dataset and derived results, DOI 10.5281/zenodo.21278557, CC BY 4.0.
- Author collection and analysis software, DOI 10.5281/zenodo.21278793, MIT.
- VerifyLLMAPI atlas JSON, built from reference
pamela-openrouter-2026-07-8cell-ref-a, SHA-25651f3f63cbcfb56b151055d2d10e36a9cc2f0644cc68b580784853d68951fa866.
Page evidence SHA-256: 6257544a6bedc053f81fb7c6c9d614cb42b4511c03ff1f37986a9ca70bae8ed4. This hash covers the model ID, eight cell metrics, nearest references, and strongest cell difference in the generated atlas.