Kimi K2

All models, with previews →

Moonshot AI · Jul 2025Latent attention and experts

Checkpoint configuration and implementation moonshotai/Kimi-K2-Instruct @ fd1984e · parameter count checked against the checkpoint’s tensor headers · How to read

Token embedding

×

learnedself.embed_tokens(input_ids)

Look up a learned feature vector for each token ID.

x0=E[I]{\htmlData{light=embedded embedding}{x_0}}=E[{\htmlData{light=token-ids tokenize}{I}}]
in
II [B batch, T positions]
out
x0x_0 [B batch, T positions, 7168 d]
parameters
1,174,405,120
Parameters

1,174,405,120 parameters per instance.

Exact per-instance accounting
ComponentCalculationParameters
Token embedding163,840×7,168163{,}840 \times 7{,}1681,174,405,120
Per-instance total1,174,405,120

Derived from the configured table dimensions; the model total matches the checkpoint’s tensors.

Codemodeling_deepseek.py 1360–1362, 1440–1441 · config.json 1–68
modeling_deepseek.py 1360–1362
        self.embed_tokens = nn.Embedding(
            config.vocab_size, config.hidden_size, self.padding_idx
        )
modeling_deepseek.py 1440–1441
        if inputs_embeds is None:
            inputs_embeds = self.embed_tokens(input_ids)
config.json 1–68
{
  "architectures": [
    "DeepseekV3ForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "auto_map": {
    "AutoConfig": "configuration_deepseek.DeepseekV3Config",
    "AutoModel": "modeling_deepseek.DeepseekV3Model",
    "AutoModelForCausalLM": "modeling_deepseek.DeepseekV3ForCausalLM"
  },
  "aux_loss_alpha": 0.001,
  "bos_token_id": 163584,
  "eos_token_id": 163585,
  "first_k_dense_replace": 1,
  "hidden_act": "silu",
  "hidden_size": 7168,
  "initializer_range": 0.02,
  "intermediate_size": 18432,
  "kv_lora_rank": 512,
  "max_position_embeddings": 131072,
  "model_type": "kimi_k2",
  "moe_intermediate_size": 2048,
  "moe_layer_freq": 1,
  "n_group": 1,
  "n_routed_experts": 384,
  "n_shared_experts": 1,
  "norm_topk_prob": true,
  "num_attention_heads": 64,
  "num_experts_per_tok": 8,
  "num_hidden_layers": 61,
  "num_key_value_heads": 64,
  "num_nextn_predict_layers": 0,
  "pretraining_tp": 1,
  "q_lora_rank": 1536,
  "qk_nope_head_dim": 128,
  "qk_rope_head_dim": 64,
  "quantization_config": {
    "activation_scheme": "dynamic",
    "fmt": "e4m3",
    "quant_method": "fp8",
    "weight_block_size": [
      128,
      128
    ]
  },
  "rms_norm_eps": 1e-06,
  "rope_theta": 50000.0,
  "routed_scaling_factor": 2.827,
  "rope_scaling": {
    "beta_fast": 1.0,
    "beta_slow": 1.0,
    "factor": 32.0,
    "mscale": 1.0,
    "mscale_all_dim": 1.0,
    "original_max_position_embeddings": 4096,
    "type": "yarn"
  },
  "scoring_func": "sigmoid",
  "seq_aux": true,
  "tie_word_embeddings": false,
  "topk_group": 1,
  "topk_method": "noaux_tc",
  "torch_dtype": "bfloat16",
  "transformers_version": "4.48.3",
  "use_cache": true,
  "v_head_dim": 128,
  "vocab_size": 163840
Sources2

Direct evidence

The token embedding is a learned table of vocabulary × width rows, looked up by token ID.

embedding · moonshotai/Kimi-K2-Instruct / modeling_deepseek.py
Pinned revision fd1984e2b7a3 ↗ · lines 1360–1441

Read code excerpt
        self.embed_tokens = nn.Embedding(
            config.vocab_size, config.hidden_size, self.padding_idx
        )

… lines 1363–1439 omitted …

        if inputs_embeds is None:
            inputs_embeds = self.embed_tokens(input_ids)

Direct evidence

The released configuration: width, layers, dense layers, heads, routed and shared experts, routing (sigmoid scores, one expert group, top-8 of 384, route scale 2.827), MLA ranks and head widths, vocabulary, YaRN RoPE.

config · moonshotai/Kimi-K2-Instruct / config.json
Pinned revision fd1984e2b7a3 ↗ · lines 1–68

Read code excerpt
{
  "architectures": [
    "DeepseekV3ForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "auto_map": {
    "AutoConfig": "configuration_deepseek.DeepseekV3Config",
    "AutoModel": "modeling_deepseek.DeepseekV3Model",
    "AutoModelForCausalLM": "modeling_deepseek.DeepseekV3ForCausalLM"
  },
  "aux_loss_alpha": 0.001,
  "bos_token_id": 163584,
  "eos_token_id": 163585,
  "first_k_dense_replace": 1,
  "hidden_act": "silu",
  "hidden_size": 7168,
  "initializer_range": 0.02,
  "intermediate_size": 18432,
  "kv_lora_rank": 512,
  "max_position_embeddings": 131072,
  "model_type": "kimi_k2",
  "moe_intermediate_size": 2048,
  "moe_layer_freq": 1,
  "n_group": 1,
  "n_routed_experts": 384,
  "n_shared_experts": 1,
  "norm_topk_prob": true,
  "num_attention_heads": 64,
  "num_experts_per_tok": 8,
  "num_hidden_layers": 61,
  "num_key_value_heads": 64,
  "num_nextn_predict_layers": 0,
  "pretraining_tp": 1,
  "q_lora_rank": 1536,
  "qk_nope_head_dim": 128,
  "qk_rope_head_dim": 64,
  "quantization_config": {
    "activation_scheme": "dynamic",
    "fmt": "e4m3",
    "quant_method": "fp8",
    "weight_block_size": [
      128,
      128
    ]
  },
  "rms_norm_eps": 1e-06,
  "rope_theta": 50000.0,
  "routed_scaling_factor": 2.827,
  "rope_scaling": {
    "beta_fast": 1.0,
    "beta_slow": 1.0,
    "factor": 32.0,
    "mscale": 1.0,
    "mscale_all_dim": 1.0,
    "original_max_position_embeddings": 4096,
    "type": "yarn"
  },
  "scoring_func": "sigmoid",
  "seq_aux": true,
  "tie_word_embeddings": false,
  "topk_group": 1,
  "topk_method": "noaux_tc",
  "torch_dtype": "bfloat16",
  "transformers_version": "4.48.3",
  "use_cache": true,
  "v_head_dim": 128,
  "vocab_size": 163840