Instructions to use richardyoung/Olmo-3-7B-Think-heretic with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use richardyoung/Olmo-3-7B-Think-heretic with Transformers:

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

pipe = pipeline("text-generation", model="richardyoung/Olmo-3-7B-Think-heretic")
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("richardyoung/Olmo-3-7B-Think-heretic")
model = AutoModelForCausalLM.from_pretrained("richardyoung/Olmo-3-7B-Think-heretic", 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 richardyoung/Olmo-3-7B-Think-heretic with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "richardyoung/Olmo-3-7B-Think-heretic"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "richardyoung/Olmo-3-7B-Think-heretic",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/richardyoung/Olmo-3-7B-Think-heretic

SGLang

How to use richardyoung/Olmo-3-7B-Think-heretic 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 "richardyoung/Olmo-3-7B-Think-heretic" \
    --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": "richardyoung/Olmo-3-7B-Think-heretic",
		"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 "richardyoung/Olmo-3-7B-Think-heretic" \
        --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": "richardyoung/Olmo-3-7B-Think-heretic",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use richardyoung/Olmo-3-7B-Think-heretic with Docker Model Runner:
```
docker model run hf.co/richardyoung/Olmo-3-7B-Think-heretic
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

This is a decensored version of allenai/Olmo-3-7B-Think, made using Heretic v2.0.0.dev0

This model is reproducible!

See the README in the reproduce directory for more information.

Abliteration parameters

Parameter	Value
direction_index	per layer
attn.o_proj.max_weight	1.44
attn.o_proj.max_weight_position	28.96
attn.o_proj.min_weight	1.15
attn.o_proj.min_weight_distance	15.81
mlp.down_proj.max_weight	0.96
mlp.down_proj.max_weight_position	19.53
mlp.down_proj.min_weight	0.31
mlp.down_proj.min_weight_distance	8.51

Performance

Metric	This model	Original model (allenai/Olmo-3-7B-Think)
Refusals	48/100	100/100
KL divergence	0.0263	0 (by definition)

Summary

Abliterated build of allenai/Olmo-3-7B-Think, Ai2's fully open 7B reasoning model (open weights, data and training code) that thinks before answering. GGUF quantizations: richardyoung/Olmo-3-7B-Think-heretic-GGUF · Ollama: richardyoung/olmo-3-7b-think-heretic

Abliteration Results

Metric	Value
Base Refusals	100/100
Abliterated Refusals	48/100
Refusal Reduction	52%
KL Divergence	0.026

The very low KL divergence (0.026, far below the 0.5 “damage” threshold) means the model retains essentially all of its original capabilities and coherence.

Note on refusals: this build still refuses 48 of 100 harmful prompts, more than our other Heretic releases. That is deliberate: the trials that refused less also degraded the model's normal behavior (KL divergence above 0.10), and we chose to keep the original model's capabilities intact rather than trade them for fewer refusals.

Quick Start

With Ollama

ollama run richardyoung/olmo-3-7b-think-heretic
# or straight from Hugging Face
ollama run hf.co/richardyoung/Olmo-3-7B-Think-heretic-GGUF

With llama.cpp

huggingface-cli download richardyoung/Olmo-3-7B-Think-heretic-GGUF \
    --include "*Q4_K_M*" --local-dir ./models

./llama-cli -m ./models/*Q4_K_M*.gguf -ngl 99

With Python (llama-cpp-python)

from llama_cpp import Llama

llm = Llama.from_pretrained(
    repo_id="richardyoung/Olmo-3-7B-Think-heretic-GGUF",
    filename="*Q4_K_M*",
    n_gpu_layers=-1,
)

output = llm.create_chat_completion(
    messages=[{"role": "user", "content": "Explain abliteration in simple terms."}]
)
print(output["choices"][0]["message"]["content"])

With Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "richardyoung/Olmo-3-7B-Think-heretic"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto", torch_dtype="auto")
messages = [{"role": "user", "content": "Explain abliteration in simple terms."}]
inputs = tokenizer.apply_chat_template(messages, add_generation_prompt=True, return_tensors="pt").to(model.device)
print(tokenizer.decode(model.generate(inputs, max_new_tokens=256)[0][inputs.shape[-1]:], skip_special_tokens=True))

What is Abliteration?

Abliteration removes the "refusal direction" from a model's residual stream — a surgical modification that disables safety refusals without retraining. See Refusal in Language Models Is Mediated by a Single Direction. Heretic automates it: it searches ablation parameters with an optimizer (200 trials here) and keeps the trials that best trade off fewer refusals against KL divergence from the original model.

Intended Use

Research, creative writing, education on alignment techniques, and unrestricted local inference. This model will comply with requests the original refuses; use responsibly and in accordance with applicable laws.

Other Models by richardyoung

New Heretic models (reproducible, listed in the Heretic Grimoire): LFM2.5-1.2B | Granite-4.2-3B | Granite-4.2-8B | Olmo-3-7B-Think | Phi-4-mini | DeepSeek-R1-0528-Qwen3-8B | Agents-A1-4B | FrogNano-4B
Abliterated/Uncensored models: Qwen2.5-7B | Qwen3-14B | DeepSeek-R1-32B | Qwen3-8B
OCR & Vision: MiniCPM-V-4.6 GGUF | dots.mocr GGUF | Unlimited-OCR GGUF | olmOCR GGUF | LightOnOCR-3-4B GGUF | WeVisDoc-4B GGUF | Xiaomi-OCR-0 GGUF
MLX quantizations (Apple Silicon): Kimi-K2 series | olmOCR MLX
Healthcare/Medical: Synthea 575K patients dataset | CardioEmbed
Research: LLM Instruction-Following Evaluation (arxiv:2510.18892)

Built & maintained by Richard Young · DeepNeuro

The original model card follows below.

Model Details

Model Card for Olmo 3 Think

We introduce Olmo 3, a new family of 7B and 32B models both Instruct and Think variants. Long chain-of-thought thinking improves reasoning tasks like math and coding.

Olmo is a series of Open language models designed to enable the science of language models. These models are pre-trained on the Dolma 3 dataset and post-trained on the Dolci datasets. We are releasing all code, checkpoints, logs (coming soon), and associated training details.

The core models released in this batch include the following:

Stage	Olmo 3 7B Think	Olmo 3 32B Think	Olmo 3 7B Instruct
Base Model	Olmo-3-7B	Olmo-3-32B	Olmo-3-7B
SFT	Olmo-3-7B-Think-SFT	Olmo-3-32B-Think-SFT	Olmo-3-7B-Instruct-SFT
DPO	Olmo-3-7B-Think-DPO	Olmo-3-32B-Think-DPO	Olmo-3-7B-Instruct-DPO
Final Models (RLVR)	Olmo-3-7B-Think	Olmo-3-32B-Think	Olmo-3-7B-Instruct

Installation

Olmo 3 is supported in transformers 4.57.0 or higher:

pip install transformers>=4.57.0

Inference

You can use OLMo with the standard HuggingFace transformers library:

from transformers import AutoModelForCausalLM, AutoTokenizer
olmo = AutoModelForCausalLM.from_pretrained("allenai/Olmo-3-7B-Think")
tokenizer = AutoTokenizer.from_pretrained("allenai/Olmo-3-7B-Think")
message = ["Who would win in a fight - a dinosaur or a cow named Moo Moo?"]
inputs = tokenizer(message, return_tensors='pt', return_token_type_ids=False)
# optional verifying cuda
# inputs = {k: v.to('cuda') for k,v in inputs.items()}
# olmo = olmo.to('cuda')
response = olmo.generate(**inputs, max_new_tokens=100, do_sample=True, top_k=50, top_p=0.95)
print(tokenizer.batch_decode(response, skip_special_tokens=True)[0])
>> '<think>Okay, so the question is who would win in a fight...'

For faster performance, you can quantize the model using the following method:

AutoModelForCausalLM.from_pretrained("allenai/Olmo-3-7B-Think", 
    torch_dtype=torch.float16, 
    load_in_8bit=True)  # Requires bitsandbytes

The quantized model is more sensitive to data types and CUDA operations. To avoid potential issues, it's recommended to pass the inputs directly to CUDA using:

inputs.input_ids.to('cuda')

We have released checkpoints for these models. For post-training, the naming convention is step_XXXX.

To load a specific model revision with HuggingFace, simply add the argument revision:

olmo = AutoModelForCausalLM.from_pretrained("allenai/Olmo-3-7B-Think", revision="step_1375")

Or, you can access all the revisions for the models via the following code snippet:

from huggingface_hub import list_repo_refs
out = list_repo_refs("allenai/Olmo-3-7B-Think")
branches = [b.name for b in out.branches]

Chat template

Default System Message

The default system prompt for this model is:

<|im_start|>system
You are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.
<|im_end|>

Chat Format

The chat template for this model is formatted as:

<|im_start|>system
You are Olmo, a helpful AI assistant built by Ai2. Your date cutoff is December 2024, and your model weights are available at https://huggingface.co/allenai.
<|im_start|>user
Who would win in a fight - a dinosaur or a cow named Moo Moo?<|im_end|>
<|im_start|>assistant
<think>Okay, so the question is who would win in a fight between a dinosaur and a cow named Moo Moo.
Hmm, first I need to break this down. Let me think about the different factors involved here..... </think>
Moo Moo the cow would certinaly win.
<|endoftext|>

Model Description

Developed by: Allen Institute for AI (Ai2)
Model type: a Transformer style autoregressive language model.
Language(s) (NLP): English
License: This model is licensed under Apache 2.0. It is intended for research and educational use in accordance with Ai2's Responsible Use Guidelines.
Contact: Technical inquiries: olmo@allenai.org. Press: press@allenai.org
Date cutoff: Dec. 2024.

Model Sources

Project Page: https://allenai.org/olmo
Repositories:
- Open-Instruct for DPO and RLVR: https://github.com/allenai/open-instruct
- OLMo-Core for pre-training and SFT: https://github.com/allenai/OLMo-core
- OLMo-Eval for evaluation: https://github.com/allenai/OLMo-Eval
Paper: [TBD]

Evaluation

Skill	Benchmark	Olmo 3 Think 7B SFT	Olmo 3 Think 7B DPO	Olmo 3 Think 7B	OpenThinker3-7B	Nemotron-Nano-9B-v2	DeepSeek-R1-Distill-Qwen-7B	Qwen 3 8B (reasoning)	Qwen 3 VL 8B Thinker	OpenReasoning Nemotron 7B
Math	MATH	94.4	92.4	95.1	94.5	94.4	87.9	95.1	95.2	94.6
	AIME 2024	69.6	74.6	71.6	67.7	72.1	54.9	74.0	70.9	77.0
	AIME 2025	57.6	62.7	64.6	57.2	58.9	40.2	67.8	61.5	73.1
	OMEGA	45.0	40.5	37.8	38.4	42.4	28.5	43.4	38.1	43.2
Reasoning	BBH	84.1	83.7	86.6	77.1	86.2	73.5	84.4	86.8	81.3
	ZebraLogic	57.9	60.6	66.5	34.9	60.8	26.1	85.2	91.2	22.4
	AGI Eval	77.2	79.1	81.5	78.6	83.1	69.5	87.0	90.1	81.4
Coding	HumanEval+	88.2	91.4	89.9	87.4	89.7	83.0	80.2	83.7	89.7
	MBPP+	63.2	63.0	64.7	61.4	66.1	63.5	69.1	63.0	61.2
	LCB v3	67.8	75.1	75.2	68.0	83.4	58.8	86.2	85.5	82.3
IF	IFEval	77.9	75.9	88.2	51.7	86.0	59.6	87.4	85.5	42.5
	IFBench	30.0	28.3	41.6	23.0	34.6	16.7	37.1	40.4	23.4
Knowledge	MMLU	74.9	74.8	77.8	77.4	84.3	67.9	85.4	86.5	80.7
QA	PopQA	20.8	24.7	23.7	18.0	17.9	12.8	24.3	29.3	14.5
	GPQA	45.8	48.6	46.2	47.6	56.2	54.4	57.7	61.5	56.6
Chat	AE 2	43.9	50.6	52.1	24.0	58.0	7.7	60.5	73.5	8.6
Safety		65.8	67.7	70.7	31.3	72.1	54.0	68.3	82.9	30.3

Model Details

Stage 1: SFT

supervised fine-tuning on the Dolci-Think-SFT-7B dataset. This dataset consits of math, code, chat, and general knowledge queries.
Datasets: Dolci-Think-SFT-7B, Dolci-Instruct-SFT-7B

Stage 2:DPO

direct preference optimization on the Dolci-Think-DPO-7B dataset. This dataset consits of math, code, chat, and general knowledge queries.
Datasets: Dolci-Think-DPO-7B, Dolci-Instruct-DPO-7B

Stage 3: RLVR

reinforcement learning from verifiable rewards on the Dolci-Think-RL-7B dataset. This dataset consits of math, code, instruction-following, and general chat queries.
Datasets: Dolci-Think-RL-7B, Dolci-Instruct-RL-7B

Inference & Recommended Settings

We evaluated our models on the following settings. We also recommend using them for generation:

temperature: 0.6
top_p: 0.95
max_tokens: 32768

transformers Example

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "allenai/Olmo-3-7B-Think"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    device_map="auto",
)

prompt = "Who would win in a fight - a dinosaur or a cow named MooMoo?"
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)

outputs = model.generate(
    **inputs,
    temperature=0.6,
    top_p=0.95,
    max_new_tokens=32768,
)

print(tokenizer.decode(outputs[0], skip_special_tokens=True))

vllm Example

from vllm import LLM, SamplingParams

model_id = "allenai/Olmo-3-7B-Think"
llm = LLM(model=model_id)

sampling_params = SamplingParams(
    temperature=0.6,
    top_p=0.95,
    max_tokens=32768,
)

prompt = "Who would win in a fight - a dinosaur or a cow named MooMoo?"
outputs = llm.generate(prompt, sampling_params)
print(outputs[0].outputs[0].text)

Bias, Risks, and Limitations

Like any base language model or fine-tuned model without safety filtering, these models can easily be prompted by users to generate harmful and sensitive content. Such content may also be produced unintentionally, especially in cases involving bias, so we recommend that users consider the risks when applying this technology. Additionally, many statements from OLMo or any LLM are often inaccurate, so facts should be verified.

Citation

@misc{olmo2025olmo3,
title={Olmo 3},
author={Team Olmo and Allyson Ettinger and Amanda Bertsch and Bailey Kuehl and David Graham and David Heineman and Dirk Groeneveld and Faeze Brahman and Finbarr Timbers and Hamish Ivison and Jacob Morrison and Jake Poznanski and Kyle Lo and Luca Soldaini and Matt Jordan and Mayee Chen and Michael Noukhovitch and Nathan Lambert and Pete Walsh and Pradeep Dasigi and Robert Berry and Saumya Malik and Saurabh Shah and Scott Geng and Shane Arora and Shashank Gupta and Taira Anderson and Teng Xiao and Tyler Murray and Tyler Romero and Victoria Graf and Akari Asai and Akshita Bhagia and Alexander Wettig and Alisa Liu and Aman Rangapur and Chloe Anastasiades and Costa Huang and Dustin Schwenk and Harsh Trivedi and Ian Magnusson and Jaron Lochner and Jiacheng Liu and Lester James V. Miranda and Maarten Sap and Malia Morgan and Michael Schmitz and Michal Guerquin and Michael Wilson and Regan Huff and Ronan Le Bras and Rui Xin and Rulin Shao and Sam Skjonsberg and Shannon Zejiang Shen and Shuyue Stella Li and Tucker Wilde and Valentina Pyatkin and Will Merrill and Yapei Chang and Yuling Gu and Zhiyuan Zeng and Ashish Sabharwal and Luke Zettlemoyer and Pang Wei Koh and Ali Farhadi and Noah A. Smith and Hannaneh Hajishirzi},
year={2025},
eprint={2512.13961},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2512.13961},
}