Instructions to use OpenGVLab/InternVL3_5-241B-A28B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use OpenGVLab/InternVL3_5-241B-A28B with Transformers:

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

pipe = pipeline("image-text-to-text", model="OpenGVLab/InternVL3_5-241B-A28B", 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 AutoModel
model = AutoModel.from_pretrained("OpenGVLab/InternVL3_5-241B-A28B", trust_remote_code=True, dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use OpenGVLab/InternVL3_5-241B-A28B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "OpenGVLab/InternVL3_5-241B-A28B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "OpenGVLab/InternVL3_5-241B-A28B",
		"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/OpenGVLab/InternVL3_5-241B-A28B

SGLang

How to use OpenGVLab/InternVL3_5-241B-A28B 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 "OpenGVLab/InternVL3_5-241B-A28B" \
    --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": "OpenGVLab/InternVL3_5-241B-A28B",
		"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 "OpenGVLab/InternVL3_5-241B-A28B" \
        --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": "OpenGVLab/InternVL3_5-241B-A28B",
		"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 OpenGVLab/InternVL3_5-241B-A28B with Docker Model Runner:
```
docker model run hf.co/OpenGVLab/InternVL3_5-241B-A28B
```

Typo in leaderboard

by iyuge2 - opened Aug 26, 2025

Discussion

iyuge2

Aug 26, 2025

Hello, thank you for your excellent work. I noticed a small typo in Table 2: the value of MMStar for GLM-4.5V should be 75.6 according to the original paper, rather than 72.9, which corresponds to GLM-4.1V-9B-Thinking. Thank you again for your valuable contribution.

LiquidAmmonia

Aug 26, 2025

Hi, thank you for your great work. We are the research team of Z.ai, we notice that in Table 2 you reported the results in LVBench [1], while the results are actually the performance of LongVideoBench [2], which is also reported in your blog's 'Video Understanding' part (https://huggingface.co/OpenGVLab/InternVL3_5-241B-A28B#video-understanding). These are two different benchmarks and we used LVBench [1] in our report.

Could you double-check whether the performance in Table 2 is from LVBench [1] or LongVideoBench [2] and solve the inconsistency in Table 2 and the 'Video Understanding' section?

[1] LVBench
@misc {wang2025lvbenchextremelongvideo,
title={LVBench: An Extreme Long Video Understanding Benchmark},
author={Weihan Wang and Zehai He and Wenyi Hong and Yean Cheng and Xiaohan Zhang and Ji Qi and Xiaotao Gu and Shiyu Huang and Bin Xu and Yuxiao Dong and Ming Ding and Jie Tang},
year={2025},
eprint={2406.08035},
archivePrefix={arXiv},
primaryClass={cs.CV},
url={https://arxiv.org/abs/2406.08035},
}

[2] LongVideoBench
@misc {wu2024longvideobenchbenchmarklongcontextinterleaved,
title={LongVideoBench: A Benchmark for Long-context Interleaved Video-Language Understanding},
author={Haoning Wu and Dongxu Li and Bei Chen and Junnan Li},
year={2024},
eprint={2407.15754},
archivePrefix={arXiv},
primaryClass={cs.CV},
url={https://arxiv.org/abs/2407.15754},
}

Weiyun1025

OpenGVLab org Aug 26, 2025

Thank you for your suggestion. We will correct these typos in the revision and release a unified update next week.

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment