Instructions to use ServiceNow-AI/SuperApriel-15B-Instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ServiceNow-AI/SuperApriel-15B-Instruct with Transformers:

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

pipe = pipeline("text-generation", model="ServiceNow-AI/SuperApriel-15B-Instruct", 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)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModelForImageTextToText
model = AutoModelForImageTextToText.from_pretrained("ServiceNow-AI/SuperApriel-15B-Instruct", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use ServiceNow-AI/SuperApriel-15B-Instruct with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ServiceNow-AI/SuperApriel-15B-Instruct"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ServiceNow-AI/SuperApriel-15B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ServiceNow-AI/SuperApriel-15B-Instruct

SGLang

How to use ServiceNow-AI/SuperApriel-15B-Instruct 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 "ServiceNow-AI/SuperApriel-15B-Instruct" \
    --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": "ServiceNow-AI/SuperApriel-15B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "ServiceNow-AI/SuperApriel-15B-Instruct" \
        --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": "ServiceNow-AI/SuperApriel-15B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use ServiceNow-AI/SuperApriel-15B-Instruct with Docker Model Runner:
```
docker model run hf.co/ServiceNow-AI/SuperApriel-15B-Instruct
```

SuperApriel-15B-Instruct

Commit History

Apr 23

02:11
report link

ostapeno

7012e86
02:09
eprint

ostapeno

8925af8

Apr 22

12:57
updated results

ostapeno

3ec808b

Apr 15

09:04
Add note about vLLM pure-recurrent placement limitation⁠

denisko +1: Claude Opus 4.6

da9a2d8
06:28
Update speedup numbers in tables from paper⁠

denisko +1: Claude Opus 4.6

9c39f06

Apr 14

07:16
Update Fast-LLM vLLM plugin branch reference⁠

denisko +1: Claude Opus 4.6

31fd523
06:36
Add vLLM serving section with OpenAI-compatible API examples⁠

denisko +1: Claude Opus 4.6

25bc346

Apr 13

07:36
Add Transformers support with AutoModelForCausalLM and preset selection⁠

denisko +1: Claude Opus 4.6

6b09478
06:40
Address reviewer feedback on README⁠

denisko +1: Claude Opus 4.6

472a888

Apr 10

16:02
Pin vLLM version to 0.15.1 (tested version)⁠

denisko +1: Claude Opus 4.6

136311d
16:02
Fix tilde rendering in vLLM section (avoid markdown strikethrough)⁠

denisko +1: Claude Opus 4.6

3f2bdcd
15:58
Add TODO placeholder for per-request preset selection API⁠

denisko +1: Claude Opus 4.6

5ef0391
15:56
Add vLLM serving instructions with preset selection and runtime switching⁠

denisko +1: Claude Opus 4.6

07563d8
13:06
Remove .DS_Store and add to .gitignore⁠

denisko +1: Claude Opus 4.6

c384df8
13:06
Add preset placement configs with paper-matching names⁠

denisko +1: Claude Opus 4.6

99684b3

Apr 8

07:10
Optimize logo image size (1.5MB -> 41KB)⁠

denisko +1: Claude Opus 4.6

89275ae
06:54
Add model card for SuperApriel-15b-Instruct⁠

denisko +1: Claude Opus 4.6

86ad99a

Apr 7

15:21
Upload folder using huggingface_hub

denisko

Verified signature
7925509
15:15
initial commit

denisko

Verified signature
41a56a9

Commit History

Apr 23

report link

eprint

Apr 22

updated results

Apr 15

Add note about vLLM pure-recurrent placement limitation⁠

Update speedup numbers in tables from paper⁠

Apr 14

Update Fast-LLM vLLM plugin branch reference⁠

Add vLLM serving section with OpenAI-compatible API examples⁠

Apr 13

Add Transformers support with AutoModelForCausalLM and preset selection⁠

Address reviewer feedback on README⁠

Apr 10

Pin vLLM version to 0.15.1 (tested version)⁠

Fix tilde rendering in vLLM section (avoid markdown strikethrough)⁠

Add TODO placeholder for per-request preset selection API⁠

Add vLLM serving instructions with preset selection and runtime switching⁠

Remove .DS_Store and add to .gitignore⁠

Add preset placement configs with paper-matching names⁠

Apr 8

Optimize logo image size (1.5MB -> 41KB)⁠

Add model card for SuperApriel-15b-Instruct⁠

Apr 7

Upload folder using huggingface_hub

initial commit