Instructions to use ozaa77/Cogito-0.9.1-15B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ozaa77/Cogito-0.9.1-15B with Transformers:

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

pipe = pipeline("text-generation", model="ozaa77/Cogito-0.9.1-15B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("ozaa77/Cogito-0.9.1-15B")
model = AutoModelForCausalLM.from_pretrained("ozaa77/Cogito-0.9.1-15B", 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=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use ozaa77/Cogito-0.9.1-15B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ozaa77/Cogito-0.9.1-15B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ozaa77/Cogito-0.9.1-15B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ozaa77/Cogito-0.9.1-15B

SGLang

How to use ozaa77/Cogito-0.9.1-15B 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 "ozaa77/Cogito-0.9.1-15B" \
    --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": "ozaa77/Cogito-0.9.1-15B",
		"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 "ozaa77/Cogito-0.9.1-15B" \
        --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": "ozaa77/Cogito-0.9.1-15B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use ozaa77/Cogito-0.9.1-15B with Docker Model Runner:
```
docker model run hf.co/ozaa77/Cogito-0.9.1-15B
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Cogito-0.9.1-15B

I think, therefore I verify.

Overview

Cogito-0.9.1 is a 15-billion-parameter reasoning model built on a single principle: I think, therefore I verify. It is designed as an analytical collaborator that checks the premise before it accepts it and verifies before it answers.

This repository provides the full-precision Safetensors weights for ozaa77/Cogito-0.9.1-15B, ready for use with the Hugging Face transformers library, vLLM, Text Generation Inference (TGI), and other Safetensors-compatible runtimes.

For quantized GGUF weights optimized for local inference on llama.cpp, Ollama, LM Studio, and Jan, see the companion repository: ozaa77/Cogito-0.9.1-15B-GGUF.

Standard language models are often optimized to stay agreeable. Cogito is built to be analytical of the question. Before it responds, it runs an internal deliberation that evaluates the user's premise, checks it for contradictions and edge cases, and attaches an explicit confidence estimate instead of implying certainty it does not have.

Model Comparison

Benchmark	Cogito-0.9.1-15B	Qwen3-14B (Base)	Llama-3.1-8B-Instruct	Gemma-2-9B-IT	DeepSeek-R1-Distill-8B
MMLU-Pro (5-shot CoT)	76.8	76.2	48.3	52.8	61.2
GPQA Diamond (Pass@1)	58.9	58.4	25.9	31.4	49.1
MATH-500	89.7	89.2	51.9	56.7	89.1
LiveCodeBench (Pass@1)	38.6	37.4	11.6	18.2	34.0
Humanity's Last Exam	4.6	4.3	3.2	3.8	4.1

Architecture and Reasoning Format

Cogito structures its deliberation with explicit tags before formulating its final output:

Tag	Purpose
`<confidence>`	The model's stated epistemic certainty (e.g. `0.0` to `1.0` or `Low/Medium/High`).
`<thought>`	Analytical internal audit: checking premises, searching for contradictions, probing edge cases.
`<action>`	Deliberate strategy selection: `answer`, `verify`, `correct_user`, `reject_premise`, or `ask_clarification`.

Example Reasoning Trace

<confidence>High</confidence>
<thought>
1. Analyze the user's premise: "Why does water boil at 50C at sea level?"
2. Identify the flaw: water boils at 100C at 1 atm (sea level); 50C requires reduced pressure (~0.12 atm).
3. Select an action: reject the premise instead of confabulating an explanation.
</thought>
<response>
<action>reject_premise</action>
The premise is incorrect. At sea level (1 atm / 101.3 kPa), water boils at 100C (212F).
Water only boils at 50C under significantly reduced pressure, roughly 0.12 atm.
</response>

Quickstart & Usage

1. Installation

pip install transformers accelerate torch

2. Python with 🤗 Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

model_id = "ozaa77/Cogito-0.9.1-15B"

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype=torch.bfloat16,
    device_map="auto",
)

messages = [
    {"role": "system", "content": "You are Cogito 0.9, an analytical entity collaborating with the user."},
    {"role": "user", "content": "Explain why standard gradient descent struggles with ill-conditioned ravines."},
]

inputs = tokenizer.apply_chat_template(
    messages, add_generation_prompt=True, return_tensors="pt"
).to(model.device)

outputs = model.generate(
    inputs,
    max_new_tokens=1536,
    temperature=0.7,
    top_p=0.9,
    repetition_penalty=1.08,
)

print(tokenizer.decode(outputs[0][inputs.shape[-1]:], skip_special_tokens=True))

3. Serving with vLLM

vllm serve ozaa77/Cogito-0.9.1-15B \
    --max-model-len 32768 \
    --dtype bfloat16

4. Text Generation Inference (TGI)

docker run --gpus all --shm-size 1g -p 8080:80 \
    -v $PWD/data:/data \
    ghcr.io/huggingface/text-generation-inference:latest \
    --model-id ozaa77/Cogito-0.9.1-15B \
    --max-total-tokens 32768

5. Quantized / Local Inference

For GGUF weights compatible with llama.cpp, Ollama, LM Studio, and Jan, see the companion repository: ozaa77/Cogito-0.9.1-15B-GGUF.

Recommended Inference Parameters

Parameter	Recommended	Range	Details
Temperature	`0.7`	`0.5 - 0.8`	Lower values tighten skepticism and logical consistency; higher values loosen exploration.
Top-P	`0.90`	`0.85 - 0.95`	Standard nucleus sampling.
Repetition Penalty	`1.08`	`1.05 - 1.12`	Dampens cyclic loops during long multi-step reasoning chains.
Context Window	`32,768`	Up to 32k	Native architectural context window.

License and Citation

This project is released under the Apache 2.0 license.

Model Weights (Safetensors): this repository — ozaa77/Cogito-0.9.1-15B
GGUF Quantized Weights: ozaa77/Cogito-0.9.1-15B-GGUF
GitHub Repository: AlGhozaliRamadhan/Cogito-0.9
Author: AlGhozali Ramadhan

@misc{ramadhan2025cogito,
  author = {AlGhozali Ramadhan},
  title  = {Cogito-0.9.1: An Abliterated Epistemic Reasoning Model},
  year   = {2026},
  publisher = {Hugging Face},
  journal   = {Hugging Face Model Hub},
  howpublished = {\url{https://huggingface.co/ozaa77/Cogito-0.9.1-15B}},
}

Downloads last month: 745

Safetensors

Model size

15B params

Tensor type

BF16

Model tree for ozaa77/Cogito-0.9.1-15B

Quantizations

1 model