---
pipeline_tag: text-generation
base_model:
- moonshotai/Kimi-K2-Thinking
license: other
license_name: modified-mit
library_name: Model Optimizer
tags:
- nvidia
- ModelOpt
- Kimi-K2
- quantized
- FP4
- fp4
---
# Model Overview

## Description:
The NVIDIA Kimi-K2-Thinking-NVFP4 model is the quantized version of the Moonshot AI's Kimi-K2-Thinking model, which is an auto-regressive language model that uses an optimized transformer architecture. For more information, please check [here](https://huggingface.co/moonshotai/Kimi-K2-Thinking). The NVIDIA Kimi-K2-Thinking NVFP4 model is quantized with [TensorRT Model Optimizer](https://github.com/NVIDIA/TensorRT-Model-Optimizer).

This model is ready for commercial/non-commercial use.  <br>

## Third-Party Community Consideration
This model is not owned or developed by NVIDIA. This model has been developed and built to a third-party’s requirements for this application and use case; see link to Non-NVIDIA [(Kimi-K2-Thinking) Model Card](https://huggingface.co/moonshotai/Kimi-K2-Thinking).

### License/Terms of Use:
[Modified MIT](https://huggingface.co/nvidia/Kimi-K2-Thinking-NVFP4/blob/main/LICENSE)

### Deployment Geography
Global

### Use Case
This model is intended for developers and researchers building LLMs

### Release Date
Huggingface 12/04/2025 via https://huggingface.co/nvidia/Kimi-K2-Thinking-NVFP4


## Model Architecture:
**Architecture Type:** Transformers  <br>
**Network Architecture:** DeepSeek V3 <br>
**Number of Model Parameters:** 1T

## Input:
**Input Type(s):** Text <br>
**Input Format(s):** String <br>
**Input Parameters:** 1D (One Dimensional): Sequences <br>

## Output:
**Output Type(s):** Text <br>
**Output Format:** String <br>
**Output Parameters:** 1D (One Dimensional): Sequences <br>

Our AI models are designed and/or optimized to run on NVIDIA GPU-accelerated systems. By leveraging NVIDIA’s hardware (e.g. GPU cores) and software frameworks (e.g., CUDA libraries), the model achieves faster training and inference times compared to CPU-only solutions.

## Software Integration:
**Supported Runtime Engine(s):** <br>
* vLLM <br>

**Supported Hardware Microarchitecture Compatibility:** <br>
* NVIDIA Blackwell <br>

**Preferred Operating System(s):** <br>
* Linux <br>
The integration of foundation and fine-tuned models into AI systems requires additional testing using use-case-specific data to ensure safe and effective deployment. Following the V-model methodology, iterative testing and validation at both unit and system levels are essential to mitigate risks, meet technical and functional requirements, and ensure compliance with safety and ethical standards before deployment.

## Model Version(s):
** The model is quantized with nvidia-modelopt **v0.39.0**  <br>

## Calibration Datasets:
* Calibration Dataset: [cnn_dailymail](https://huggingface.co/datasets/abisee/cnn_dailymail) <br>
** Data collection method: Automated. <br>
** Labeling method: Automated. <br>

## Training Dataset: <br>
** Data Collection Method by dataset: Hybrid: Human, Automated <br>
** Labeling Method by dataset: Hybrid: Human, Automated <br>
** Data Modality: [Text] <br>
** Text Training Data Size: undisclosed. <br>

## Testing Dataset: <br>
** Data Collection Method by dataset: Hybrid: Human, Automated <br>
** Labeling Method by dataset: Hybrid: Human, Automated <br>

## Evaluation Dataset: <br>
** Data Collection Method by dataset: Hybrid: Human, Automated <br>
** Labeling Method by dataset: Hybrid: Human, Automated <br>


## Inference:
**Engine:** vLLM/SGLang/TensorRT-LLM <br>
**Test Hardware:** B200 <br>

## Post Training Quantization
This model was obtained by converting and quantizing the weights and activations of Kimi-K2-Thinking from INT4 to BF16 to NVFP4 data type, ready for inference with vLLM. Only the weights and activations of the linear operators within transformer blocks in MoE are quantized.

## Usage


To serve this checkpoint with [vLLM](https://github.com/vllm-project/vllm), you can start the docker `vllm/vllm-openai:v0.11.2` and run the sample command below:

```sh
python3 -m vllm.entrypoints.openai.api_server --model nvidia/Kimi-K2-Thinking-NVFP4 --trust-remote-code --tensor-parallel-size 4
```

To serve this checkpoint with [SGLang](https://github.com/sgl-project/sglang), you can start the docker `lmsysorg/sglang:latest` and run the sample command below:

```sh
python3 -m sglang.launch_server --model-path nvidia/Kimi-K2-Thinking-NVFP4 --tp 4 --quantization modelopt_fp4 --trust-remote-code
```

To serve this checkpoint with [TensorRT-LLM](https://github.com/NVIDIA/TensorRT-LLM), please follow instructions in [deployment-guide-for-kimi-k2-thinking-on-trtllm](https://github.com/NVIDIA/TensorRT-LLM/blob/main/docs/source/deployment-guide/deployment-guide-for-kimi-k2-thinking-on-trtllm.md)


## Model Limitations:
The base model was trained on data that contains toxic language and societal biases originally crawled from the internet. Therefore, the model may amplify those biases and return toxic responses especially when prompted with toxic prompts. The model may generate answers that may be inaccurate, omit key information, or include irrelevant or redundant text producing socially unacceptable or undesirable text, even if the prompt itself does not include anything explicitly offensive.

## Ethical Considerations

NVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications.  When downloaded or used in accordance with our terms of service, developers should work with their internal model team to ensure this model meets requirements for the relevant industry and use case and addresses unforeseen product misuse.

Please report security vulnerabilities or NVIDIA AI Concerns [here](https://www.nvidia.com/en-us/support/submit-security-vulnerability/).