Kimi K3

All models, with previews →

Moonshot AI · Release date not pinnedNew parts in the block · after Kimi K2 in the atlas

Checkpoint configuration and implementation moonshotai/Kimi-K3 @ f831ab6, fla-org/flash-linear-attention @ 07ca1e4 · parameter count checked against the checkpoint’s tensor headers · text model only · How to read

Full sequencePrefillDecode

Checkpoint tokenizer

×

no parametersTikTokenTokenizer

The tiktoken vocabulary and tokenizer class shipped with the checkpoint.

Explanation

The checkpoint’s own TikTokenTokenizer class loads the vocabulary; tokenizer_config.json names it. The vocabulary file is not among the pinned sources.

Codetokenizer_config.json 149–156
tokenizer_config.json 149–156
  "tokenizer_class": "TikTokenTokenizer",
  "unk_token": "[UNK]",
  "auto_map": {
    "AutoTokenizer": [
      "tokenization_kimi.TikTokenTokenizer",
      null
    ]
  }
Sources1

Contextual evidence

The checkpoint’s tokenizer is the repository’s own TikTokenTokenizer class, loaded through AutoTokenizer.

tokenizer · moonshotai/Kimi-K3 / tokenizer_config.json
Pinned revision f831ab668142 ↗ · lines 149–156

Read code excerpt
  "tokenizer_class": "TikTokenTokenizer",
  "unk_token": "[UNK]",
  "auto_map": {
    "AutoTokenizer": [
      "tokenization_kimi.TikTokenTokenizer",
      null
    ]
  }