Instructions to use BlackfrostAI/Kimi-K2.7-Code-BF16 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use BlackfrostAI/Kimi-K2.7-Code-BF16 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("image-text-to-text", model="BlackfrostAI/Kimi-K2.7-Code-BF16", trust_remote_code=True)
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# Load model directly
from transformers import AutoProcessor, AutoModel

processor = AutoProcessor.from_pretrained("BlackfrostAI/Kimi-K2.7-Code-BF16", trust_remote_code=True)
model = AutoModel.from_pretrained("BlackfrostAI/Kimi-K2.7-Code-BF16", trust_remote_code=True, device_map="auto")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use BlackfrostAI/Kimi-K2.7-Code-BF16 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "BlackfrostAI/Kimi-K2.7-Code-BF16"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "BlackfrostAI/Kimi-K2.7-Code-BF16",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/BlackfrostAI/Kimi-K2.7-Code-BF16

SGLang

How to use BlackfrostAI/Kimi-K2.7-Code-BF16 with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "BlackfrostAI/Kimi-K2.7-Code-BF16" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "BlackfrostAI/Kimi-K2.7-Code-BF16",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "BlackfrostAI/Kimi-K2.7-Code-BF16" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "BlackfrostAI/Kimi-K2.7-Code-BF16",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use BlackfrostAI/Kimi-K2.7-Code-BF16 with Docker Model Runner:
```
docker model run hf.co/BlackfrostAI/Kimi-K2.7-Code-BF16
```

ℹ️ Full BF16 Weights (dequantized)

This repository provides full bfloat16 weights for moonshotai/Kimi-K2.7-Code, which Moonshot AI released in native INT4 (pack-quantized / compressed-tensors) format.

The weights here are a faithful dequantization of the official INT4 release to bfloat16 — no fine-tuning, ablation, distillation, or other behavioral modification. quantization_config has been removed from config.json and dtype set to bfloat16, so the model loads as a standard full-precision transformers model.

Base model moonshotai/Kimi-K2.7-Code (© Moonshot AI, Modified MIT)

This repo BF16 dequantization by Blackfrost-AI

Precision bfloat16 (attention, shared experts, and routed experts all upcast from INT4)

Size ~2.05 TB · 64 safetensors shards · 70,310 tensors

Behavior Identical to the original — all architecture, capabilities, and evaluations below are Moonshot's

Intended use A full-precision base for re-quantization (GGUF / NVFP4 / FP8), fine-tuning, or research that needs unpacked weights

All credit for the model belongs to Moonshot AI. This conversion is provided under the same Modified MIT License as the original.


Base model	`moonshotai/Kimi-K2.7-Code` (© Moonshot AI, Modified MIT)
This repo	BF16 dequantization by Blackfrost-AI
Precision	`bfloat16` (attention, shared experts, and routed experts all upcast from INT4)
Size	~2.05 TB · 64 safetensors shards · 70,310 tensors
Behavior	Identical to the original — all architecture, capabilities, and evaluations below are Moonshot's
Intended use	A full-precision base for re-quantization (GGUF / NVFP4 / FP8), fine-tuning, or research that needs unpacked weights

1. Model Introduction

Kimi K2.7 Code is a coding-focused agentic model built upon Kimi K2.6. With substantial improvements on real-world long-horizon coding tasks, it strengthens end-to-end task completion across complex software engineering workflows while improving token efficiency, reducing thinking-token usage by approximately 30% compared with Kimi K2.6.

2. Model Summary


Architecture	Mixture-of-Experts (MoE)
Total Parameters	1T
Activated Parameters	32B
Number of Layers (Dense layer included)	61
Number of Dense Layers	1
Attention Hidden Dimension	7168
MoE Hidden Dimension (per Expert)	2048
Number of Attention Heads	64
Number of Experts	384
Selected Experts per Token	8
Number of Shared Experts	1
Vocabulary Size	160K
Context Length	256K
Attention Mechanism	MLA
Activation Function	SwiGLU
Vision Encoder	MoonViT
Parameters of Vision Encoder	400M
Weight Precision (this repo)	BF16 (dequantized from native INT4)

3. Evaluation Results

Evaluations are those of the original moonshotai/Kimi-K2.7-Code. Dequantization to BF16 does not change model behavior; if anything it removes INT4 rounding.

Benchmark	Kimi K2.6	Kimi K2.7 Code	GPT-5.5	Claude Opus 4.8
Coding
Kimi Code Bench v2	50.9	62.0	69.0	67.4
Program Bench	48.3	53.6	69.1	63.8
MLS Bench Lite	26.7	35.1	35.5	42.8
Agentic
Kimi Claw 24/7 Bench	42.9	46.9	52.8	50.4
MCP Atlas	69.4	76.0	79.4	81.3
MCP Mark Verified	72.8	81.1	92.9	76.4

4. Quantization / Precision

The original Kimi-K2.7-Code is released in native INT4 (pack-quantized), using the same method as Kimi-K2-Thinking.

This repository is the full BF16 dequantization of that release. The INT4 packed routed-expert weights were unpacked and dequantized to bfloat16; attention, shared experts, dense layer 0, embeddings, lm_head, and the vision tower (already BF16 in the original) are carried through unchanged. config.json has quantization_config removed and dtype: bfloat16, so it loads as a standard BF16 model — no compressed-tensors needed.

Use these weights as a full-precision starting point to re-quantize (GGUF, NVFP4, FP8, AWQ, …) or to fine-tune.

5. Deployment

Kimi-K2.7-Code has the same architecture as Kimi-K2.5/Kimi-K2.6, and the deployment method can be directly reused. Recommended inference engines:

vLLM
SGLang
KTransformers

This BF16 checkpoint loads with trust_remote_code=True at dtype=torch.bfloat16. Note it is ~2 TB — plan tensor-parallel/offload accordingly. For the original INT4 release and its official deployment guide, see moonshotai/Kimi-K2.7-Code.

6. Model Usage

Usage is identical to the original model (forced thinking + preserve_thinking). See the original model card for chat-completion, vision, preserve-thinking, and tool-call examples. The chat template and tokenizer in this repo are the originals, unchanged.

7. License

8. Third Party Notices

See THIRD PARTY NOTICES.

9. Acknowledgements & Contact

Full credit to Moonshot AI for creating and open-sourcing Kimi-K2.7-Code. This repository only redistributes a dequantized copy of their weights. For questions about the original model, contact support@moonshot.ai.

Downloads last month: 22

Safetensors

Model size

1T params

Tensor type

BF16

F32

Model tree for BlackfrostAI/Kimi-K2.7-Code-BF16

Base model

moonshotai/Kimi-K2.7-Code

Finetuned

(9)

this model