Instructions to use Scinetics-ai/ScienceRSI-8b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Scinetics-ai/ScienceRSI-8b with Transformers:

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

pipe = pipeline("text-generation", model="Scinetics-ai/ScienceRSI-8b")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Scinetics-ai/ScienceRSI-8b")
model = AutoModelForCausalLM.from_pretrained("Scinetics-ai/ScienceRSI-8b", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Scinetics-ai/ScienceRSI-8b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Scinetics-ai/ScienceRSI-8b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Scinetics-ai/ScienceRSI-8b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Scinetics-ai/ScienceRSI-8b

SGLang

How to use Scinetics-ai/ScienceRSI-8b 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 "Scinetics-ai/ScienceRSI-8b" \
    --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": "Scinetics-ai/ScienceRSI-8b",
		"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 "Scinetics-ai/ScienceRSI-8b" \
        --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": "Scinetics-ai/ScienceRSI-8b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Scinetics-ai/ScienceRSI-8b with Docker Model Runner:
```
docker model run hf.co/Scinetics-ai/ScienceRSI-8b
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

ScienceRSI-8b

ScienceRSI-8b is an 8-billion-parameter scientific reasoning model released by Scinetics.

ScienceRSI explores how scientific capabilities can become reusable foundations for further model improvement. Its central idea is a recurring cycle:

Specialize: develop complementary scientific capabilities.
Verify: identify useful contributions through shared scientific evaluation.
Transfer: integrate verified contributions into a target model.
Reuse: use the improved model as the starting point for further specialization and integration.

This release provides an 8B checkpoint for exploring that approach. The framework's long-term goal is cumulative improvement across model generations; this model card does not claim that repeated recursive rounds have already been demonstrated by this checkpoint.

Model information

Item	Value
Release name	ScienceRSI-8b
Parameters	8,190,735,360
Architecture	Qwen3ForCausalLM
Weight format	BF16 Safetensors, four shards
Configured context length	40,960 tokens
License	Apache-2.0

Load the model

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "Scinetics-ai/ScienceRSI-8b"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype="auto",
    device_map="auto",
)

Use the included tokenizer and chat template. The root directory contains the complete model; lora_adapter/ is preserved as a separate source artifact and is not required to load the root checkpoint.

Intended use

Scientific reasoning research, reproducible evaluation, and further study of reusable scientific capabilities. Generated answers can contain errors; check calculations, assumptions, and evidence independently.

Release provenance

This release renames and republishes the existing checkpoint. Model weights, tokenizer assets, configuration, and adapter files are preserved without retraining or numerical conversion. Original repository history is retained, and the existing Apache-2.0 license declaration is preserved.

Source revision: f1d0ca420a820cb6bf4ad083bb21cda5322c0c42.

Downloads last month: 32

Safetensors

Model size

8B params

Tensor type

BF16

Model tree for Scinetics-ai/ScienceRSI-8b

Quantizations

2 models