moonshotai/Kimi-K2-Thinking

moonshotai/Kimi-K2-Thinking is listed on Hugging Face as a text-generation model. This page records the public repository snapshot a51ccc050d73dab088bf7b0e2dd9b30ae85a4e55; verify the listed revision, licence, and runtime support before using it.

Updated

Computraq is preparing for launch. These guides explain readiness and the proposed framework; they do not establish a live service, integration or partnership.

Identity and access

The publisher record identifies this repository as moonshotai/Kimi-K2-Thinking. Its current public snapshot is a51ccc050d73dab088bf7b0e2dd9b30ae85a4e55. Access is public repository; review the applicable licence and publisher terms; the recorded licence label is other. A public repository or gated download does not establish permission for every use.

Configuration facts

The pinned configuration reports model type kimi_k2 and architecture DeepseekV3ForCausalLM. The model hub tags it for text-generation; these labels describe repository metadata and do not establish benchmark performance.

  • hidden_size: 7168
  • intermediate_size: 18432
  • max_position_embeddings: 262144
  • n_routed_experts: 384
  • num_attention_heads: 64
  • num_experts_per_tok: 8
  • num_hidden_layers: 61
  • num_key_value_heads: 64
  • quantization_config.config_groups.group_0.weights.group_size: 32
  • quantization_config.format: pack-quantized
  • quantization_config.quant_method: compressed-tensors
  • rms_norm_eps: 1e-05
  • rope_theta: 50000.0
  • torch_dtype: bfloat16
  • vocab_size: 163840
  • quantization_config.quantization_status: compressed

Model-specific qualification

This configuration or publisher metadata marks the model as a mixture of experts. Confirm whether the serving runtime supports its routing layers, and distinguish total parameters from parameters active per token; no active count is inferred here unless the publisher explicitly states one.

This repository appears to describe a quantized or reduced-precision variant (not stated in the public record). Confirm the actual weight format and runtime kernels from the pinned files; storage size alone is not a logical parameter count or a guaranteed GPU-memory requirement.

Weight-memory planning

This guide does not retain an independently verified logical parameter count, so it omits a numeric weight-memory estimate. If a count is verified for this exact repository revision, use parameter count × bytes per parameter for a weights-only floor; then account separately for KV cache, context, activations, runtime, and concurrency.

Revision and reproducibility

The metadata and configuration are pinned to revision a51ccc050d73dab088bf7b0e2dd9b30ae85a4e55. Pin this commit when reproducing an evaluation. Repository branches can change, and this source check does not validate the model weights themselves.

Published record

  • Task: text-generation
  • Model type: kimi_k2
  • Architecture: DeepseekV3ForCausalLM
  • Licence label: other
  • Access: public repository; review the applicable licence and publisher terms
  • Revision: a51ccc050d73dab088bf7b0e2dd9b30ae85a4e55

Sources and verification