AI Desk · AI Fingerprinting · Compare
MoonshotAI: Kimi K3 · Tencent: Hy-MT2-1.8B
Identity similarity: lowDivergent fingerprints · 35/50These models tokenise differently. The rows that part here are the English prose, whitespace, CJK, non-Latin script and emoji strings. Divergence on high-information strings separates unrelated vocabularies within a handful of samples.
MoonshotAI: Kimi K3 recordTencent: Hy-MT2-1.8B recordOpen in the interactive bench
Tokenizer fingerprint
35 of 50 mutually clean strings agree| Test string | MoonshotAI: Kimi K3 | Tencent: Hy-MT2-1.8B | Verdict |
|---|---|---|---|
| en-prose | 14 | 14 | match |
| en-long | 16 | 17 | differs |
| spaces-20 | 3 | 3 | match |
| spaces-60 | 3 | 3 | match |
| tabs-20 | 4 | 4 | match |
| newlines-20 | 4 | 4 | match |
| mixed-ws | 6 | 7 | differs |
| digits-9 | 3 | 3 | match |
| digits-12 | 4 | 4 | match |
| digits-30 | 10 | 10 | match |
| digits-sep | 7 | 7 | match |
| float-long | 9 | 9 | match |
| zh-common | 9 | 8 | differs |
| zh-long | 10 | 10 | match |
| zh-rare | 18 | 18 | match |
| ja-kana | 11 | 11 | match |
| ja-kanji | 14 | 12 | differs |
| ko | 13 | 13 | match |
| ru | 16 | 16 | match |
| ar | 15 | 15 | match |
| he | 16 | 16 | match |
| hi | 17 | 24 | differs |
| th | 21 | 17 | differs |
| el | 22 | 22 | match |
| emoji-basic | 10 | 10 | match |
| emoji-skin | 20 | 25 | differs |
| emoji-zwj-family | 11 | 11 | match |
| emoji-zwj-x3 | 33 | 33 | match |
| emoji-flags | 14 | 24 | differs |
| emoji-prof | 20 | 23 | differs |
| math | 26 | 26 | match |
| boxdraw | 24 | 30 | differs |
| combining | 14 | 10 | differs |
| cjk-ext-b | 16 | 16 | match |
| surrogates | 31 | 40 | differs |
| zalgo | 36 | 36 | match |
| rtl-mix | 8 | 8 | match |
| py-code | 25 | 25 | match |
| py-indent | 15 | 15 | match |
| json | 20 | 20 | match |
| html | 16 | 16 | match |
| regex | 43 | 45 | differs |
| camel | 6 | 6 | match |
| snake | 6 | 6 | match |
| rare-word-x5 | 31 | 31 | match |
| repeat-tok | 41 | 41 | match |
| base64 | 33 | 35 | differs |
| hex | 11 | 12 | differs |
| url | 17 | 17 | match |
| uuid | 27 | 27 | match |
Values are the marginal prompt-token cost of each string with the model's fixed overhead subtracted. An x marks a row excluded as corrupt for that model.
API surface
The declared serving contracts differ| Field | MoonshotAI: Kimi K3 | Tencent: Hy-MT2-1.8B | Verdict |
|---|---|---|---|
| Context length | 1048576 | 8192 | × differs |
| Max output tokens | – | 4096 | × differs |
| Supported parameters | frequency_penalty, include_reasoning, logit_bias, logprobs, max_tokens, min_p, presence_penalty, reasoning, reasoning_effort, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_p | max_completion_tokens, max_tokens, stop, temperature | × differs |
| Defaults | {"temperature":null,"top_p":0.95,"top_k":null,"frequency_penalty":null,"presence_penalty":null,"repetition_penalty":null} | {} | × differs |
| Reasoning contract | {"mandatory":false,"default_enabled":true,"supported_efforts":["max","high","low"],"default_effort":"max"} | – | × differs |
| Modalities | text+image+video->text | text->text | × differs |
MoonshotAI: Kimi K3: OpenRouter's listing shows 943,718 output tokens, which is 90% of the context window, the figure it carries where the provider declares no maximum.
Method: tokenizer fingerprinting, v1, 50 strings. This page is static and dated; the interactive bench compares any two of the 338 measured models.