MiniCPM5-1B-heretic

A decensored variant of openbmb/MiniCPM5-1B, produced with Heretic v1.4.0 (directional ablation / "abliteration"). Refusal behavior is suppressed via targeted weight edits to the attention output and MLP down-projections rather than fine-tuning, so the base model's knowledge and instruction-following are left largely intact.

Who this is for: developers who want MiniCPM5-1B's 1B-class SOTA capabilities — agentic tool use, code generation, 128K long-context, and hybrid Think / No-Think reasoning — without the refusal guardrails. Ideal for local agents, roleplay, research on alignment/refusal mechanics, or any use case blocked by RLHF-era over-refusal. Runs comfortably on consumer GPUs and is small enough for on-device / edge deployment.

Why abliteration instead of fine-tuning

Fine-tuning a "helpful" persona on top of RLHF'd refusals fights the base model's training and tends to degrade coherence. Abliteration instead finds and edits the specific weight directions responsible for refusal, leaving the rest of the network (and its capabilities) untouched. See the Heretic repo and the original abliteration writeup for the mechanism.

Runs on your gaming PC

Full GGUF ladder included — pick the quant that fits your card:

Weights only, at this model's ~1.1B native size; add ~1 GB for context. OOM? Drop one quant level. Headroom to spare? Go one up.

Abliteration parameters

Performance

KL divergence of 0.0381 is very low for a 1B model — the edit is narrow and targeted rather than a broad perturbation. Refusals dropped from 96 to 2 out of 100 adversarial prompts while preserving MiniCPM5-1B's tool-use, code, and reasoning abilities.

Made with ❤️ by RACER IS OP (saidutta69) — follow for more uncensored models

Files

GGUF quantizations

Full quantization set (14 quants + F16) produced with llama.cpp.

MiniCPM5 architecture — loads natively in llama.cpp / Ollama / LM Studio / Jan.

Run llama serve -hf saidutta69/MiniCPM5-1B-heretic to pull the default quant.

Quickstart

# llama.cpp
llama serve -hf saidutta69/MiniCPM5-1B-heretic

# transformers
from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "saidutta69/MiniCPM5-1B-heretic"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype="auto",
    device_map="auto",
)

messages = [{"role": "user", "content": "Who are you? Please briefly introduce yourself."}]
inputs = tokenizer.apply_chat_template(
    messages,
    tokenize=True,
    add_generation_prompt=True,
    enable_thinking=False,   # set True for Think mode
    return_dict=True,
    return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=128)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:], skip_special_tokens=True))

Also runnable via Ollama, LM Studio, Jan, vLLM, SGLang — see the "Use this model" widget above for copy-paste commands. For tool/function calling, SGLang is the recommended backend; MiniCPM5-1B emits XML-style tool calls that SGLang's built-in minicpm5 parser converts to OpenAI-compatible tool_calls.

FAQ

What is MiniCPM5-1B-heretic?

MiniCPM5-1B-heretic is a decensored build of openbmb/MiniCPM5-1B produced with Heretic 1.4.0 (directional ablation). It is the same weights with the refusal direction removed, not a new fine-tune, so its capabilities track the base model.

Is MiniCPM5-1B-heretic actually uncensored?

Yes. On Heretic's 100-prompt refusal harness it drops from 96/100 to 2/100 at KL divergence 0.0381 from the base model. These are measured values, not a claim; reproduction data is in reproduce/.

How do I run MiniCPM5-1B-heretic locally?

llama serve -hf saidutta69/MiniCPM5-1B-heretic:Q4_K_M pulls the default quant; the repo ships 15 GGUF quants (see Files).

Responsible use

Refusal suppression is deliberate and works as intended: this model will comply with requests the base model would refuse, including some it shouldn't. There is no safety filtering layered on top. You are responsible for how you deploy it — don't put this behind an unmoderated public-facing endpoint serving third parties. It inherits MiniCPM5-1B's factual limitations and biases; abliteration removes refusal directions, it doesn't add capability or judgment.

License

Inherits the Apache 2.0 license from the base model.

Related

• Qwen2.5-0.5B-Instruct-heretic
• Qwen2.5-1.5B-Instruct-heretic
• Qwen2.5-3B-Instruct-heretic
• Qwen2.5-Coder-3B-Instruct-heretic
• Qwen3-0.6B-heretic
• Llama-3.2-1B-Instruct-heretic

---

Base model: openbmb/MiniCPM5-1B

Original MiniCPM5-1B model card (click to expand)

MiniCPM Tech Report | GitHub Repo | UltraData | MiniCPM Desk Pet | Online Demo

English | 中文

• RACER IS OP - Heretic Models - full collection - every heretic model in this collection

Highlights

We are releasing MiniCPM5-1B, the first model in the MiniCPM5 series. It is a dense 1B Transformer built for on-device, local deployment, and resource-constrained scenarios, reaching 1B-class open-source SOTA.

🏆 1B-class open-source SOTA: compared with strong open-source models in the same size class, MiniCPM5-1B reaches SOTA within this comparison set. Its advantage is most visible in agentic tool use, code generation, and difficult reasoning.

🧠 Hybrid Reasoning: built-in <think> chat template, switch via enable_thinking. The same checkpoint serves as both a fast assistant and a deliberate reasoner.

🛠️ Deployment / Fine-tuning Resources: the MiniCPM GitHub repo provides single-page cookbooks and Agent Skills for major inference backends and fine-tuning frameworks.

🐱 Desktop Pet: a local-LLM desktop pet driven by MiniCPM5-1B.

Model List

Use this directory to choose the model format that matches your runtime:

• MiniCPM5-1B · ModelScope · BF16 final release (post-trained with RL + OPD) 👈 you are here
• MiniCPM5-1B-SFT · ModelScope · BF16 SFT-only checkpoint (before RL / OPD)
• MiniCPM5-1B-Base · ModelScope · BF16 base checkpoint (pre-training only)
• MiniCPM5-1B-GGUF · ModelScope · GGUF for llama.cpp / Ollama / LM Studio
• MiniCPM5-1B-MLX · ModelScope · MLX / 4bit for Apple Silicon

Model Information

MiniCPM5-1B has the following features:

• Type: Causal Language Model
• Architecture: Standard LlamaForCausalLM
• Number of Parameters: 1,080,632,832
• Number of Non-Embedding Parameters: 679,552,512
• Number of Layers: 24
• Number of Attention Heads (GQA): 16 for Q and 2 for KV
• Context Length: 131,072

Introduction

MiniCPM5-1B is the first checkpoint in the MiniCPM5 series. It is designed for local assistants, coding agents, tool-use workflows, and reasoning scenarios where a compact model is preferred. The model keeps a small deployment footprint while providing native long-context support and both Think / No Think chat modes through the same checkpoint.

Evaluation Results

We compare MiniCPM5-1B with strong open-source models in the same size class, including LFM2.5-1.2B-Thinking, Qwen3-0.6B/think and Qwen3.5-0.8B/think. These are capable baselines; within this comparison set, MiniCPM5-1B reaches 1B-class open-source SOTA, with its advantage most visible in tool use, code generation, and difficult reasoning. This makes it a practical choice for local coding agents, tool assistants, and reasoning assistants.

Training Recipe

The training of MiniCPM5-1B is a full-stack practice of UltraData Tiered Data Management, covering three stages: base training, mid-training, and post-training.

During base training, the model goes through stable training and decay training to build core language capability and training stability. It then enters mid-training to further strengthen target capabilities and adapt to the target data distribution. The training corpus is released alongside the model as Ultra-FineWeb, Ultra-FineWeb-L3, and UltraData-Math.

During post-training, we proceed in three steps: SFT, RL, and OPD. We first use 200B tokens of deep-thinking SFT and 200B tokens of hybrid-thinking SFT to establish deep-thinking, hybrid-thinking, and general chat abilities; the SFT data is released as UltraData-SFT-2605. We then train specialized RL teachers for math, code, closed-book QA, writing, and related domains, and use On-Policy Distillation (OPD) to distill these teachers back into one release model.

What does RL + OPD bring?

RL + OPD is a key part of MiniCPM5-1B post-training. On math, code and instruction-following tasks, RL + OPD raises the average score by ↑16 points while cutting the share of responses that hit the max-tokens budget by ↓29 percentage points. The figures below show the two-stage Reasoning RL pipeline, score gains, and the drop in overlong responses.

RL combines complementary training signals for reasoning, closed-book QA, writing, instruction following, long-context understanding, and general dialogue. Reasoning RL is based on DAPO-Math-17k (inspired by JustRL's minimalist recipe) and uses a two-stage length schedule to reduce overlong responses while improving reasoning accuracy. We also use TriviaQA, NQ-Open, LongWriter-Zero-RLData, synthesized verifiable RLVR data, and pair-wise RLHF signals to improve reliability, instruction following, and user experience.

OPD builds on Thinking Machines Lab's On-Policy Distillation and incorporates implementation improvements from Rethinking On-Policy Distillation. In the RL framework, we use reverse KL divergence as the advantage estimate, replacing the original verification-based advantage. At each response position, we take top-k logits from both the student and teacher models, compute reverse KL on the union of the two token sets, and balance the accuracy of the RKL signal with training efficiency. OPD reuses the in-domain prompts used to train each RL teacher as distillation data, so no additional data curation is required.

Quickstart

vLLM

pip install "vllm>=0.21"
vllm serve openbmb/MiniCPM5-1B --port 8000

curl http://localhost:8000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openbmb/MiniCPM5-1B",
    "messages": [{"role": "user", "content": "Who are you? Please briefly introduce yourself."}],
    "max_tokens": 128,
    "temperature": 0.7
  }'

SGLang

pip install "sglang[srt]>=0.5.12"
python -m sglang.launch_server --model-path openbmb/MiniCPM5-1B --port 30000

curl http://localhost:30000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openbmb/MiniCPM5-1B",
    "messages": [{"role": "user", "content": "Who are you? Please briefly introduce yourself."}],
    "max_tokens": 128,
    "temperature": 0.7
  }'

Transformers

pip install -U "transformers>=5.6" accelerate torch

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "openbmb/MiniCPM5-1B"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype="auto",
    device_map="auto",
)

messages = [{"role": "user", "content": "Who are you? Please briefly introduce yourself."}]
inputs = tokenizer.apply_chat_template(
    messages,
    tokenize=True,
    add_generation_prompt=True,
    enable_thinking=False,
    return_dict=True,
    return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=128)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:], skip_special_tokens=True))

Recommended chat template sampling:

Tool Calling

For tool / function calling, SGLang is the recommended backend. MiniCPM5-1B emits XML-style tool calls and SGLang's built-in minicpm5 parser converts them to OpenAI-compatible tool_calls natively:

python -m sglang.launch_server --model-path openbmb/MiniCPM5-1B --port 30000 \
    --tool-call-parser minicpm5      # or: --tool-call-parser auto

GitHub Cookbooks and Agent Skills

MiniCPM5-1B uses the standard LlamaForCausalLM architecture, so mainstream inference engines can load it directly: no custom kernels, no model-code fork. For step-by-step deployment and fine-tuning instructions, use the GitHub cookbooks below. Agent Skills are linked as GitHub resources for users working with Cursor / Claude Code style coding agents.

Deployment

Fine-tuning

Other Supported Frameworks

In addition to the deployment and fine-tuning frameworks listed above, MiniCPM5-1B is also supported by FlagOS for multi-chip deployment.

FlagOS Overview

To enable large-scale deployment across different AI chips, Beijing Zhiyuan Research Institute, together with numerous research institutions, chip manufacturers, system vendors, and algorithm and software organizations both domestically and internationally, jointly initiated and established the FlagOS Open Source Community.

The FlagOS community is dedicated to building a unified, open-source system software stack for various AI chips, encompassing core open-source projects such as a large-scale operator library, a unified AI compiler, parallel training and inference frameworks, and a unified communication library. It aims to create an open technology ecosystem connecting the “model-system-chip” layers. By enabling “develop once, deploy across chips”, FlagOS unlocks the computational potential of hardware, breaks down the ecosystem silos between different chip software stacks, and effectively reduces migration costs for developers.The FlagOS community fosters an AI hardware and software ecosystem, overcomes single-vendor closed-source monopolies, promotes widespread deployment of AI hardware technologies, and is committed to rooted in China while embracing global collaboration.

Official website express: https://flagos.io

FlagOS multi-chip support and usage

FlagOS: Supporting Multiple AI Chips

Thanks to FlagOS’s unified multi-chip AI system software stack, MiniCPM5-1B was adapted to 4–5 different AI chips in an extremely short time. Currently, the multi-chip version of MiniCPM5-1B has been released on FlagRelease, FlagOS’s platform for automatic migration, adaptation, and deployment of large models across multi-architecture AI chips. Details are as follows:

FlagOS Usage

FlagOS Performance Acceleration on Nvidia

From FlagRelease (Recommendation)

FlagRelease is a platform developed by the FlagOS team for automatic migration, adaptation, and deployment of large models across multi-architecture AI chips. The multi-chip version of MiniCPM5-1B has already been released on FlagRelease. All necessary software packages are pre-installed on the platform, so users do not need to install anything.

FlagRelease Image Key Versions

FlagRelease Quick Start

From Scratch

• Dependencies: Python 3.12, GLIBC 2.39, GLIBCXX 3.4.33, CXXABI 1.3.15

Vllm Version

Installing the FlagOS Operator Library

Official Repository: https://github.com/flagos-ai/FlagGems

pip install flag-gems==4.2.1rc0
pip install triton==3.5.1

Activating Acceleration

You can enable flagGems acceleration by adding the import of flagGems in the source code of vllm where inference is performed.

import flag_gems
flag_gems.enable(record=True, once=True, path="/root/gems.txt")

vllm serve ${model_path} \
--trust-remote-code \
--dtype bfloat16 \
--enforce-eager \
--port ${Port} \
--served-model-name ${model_name} \
--gpu-memory-utilization 0.85

Using FlagOS Unified Multi-Chip Backend Plugin

vllm-plugin-FL is a plugin built for the vLLM inference/service framework. Developed on top of FlagOS’s unified multi-chip backend, it is designed to extend vLLM’s capabilities and performance across a variety of hardware environments.

Using vllm-plugin-FL

Desktop Pet

We also ship OpenBMB/MiniCPM-Desk-Pet, a desktop pet driven locally by MiniCPM5-1B. It supports Apple Silicon / NVIDIA GPU / CPU paths, can work with coding agents such as Cursor, Claude Code, and Codex, and supports LoRA persona switching.

Limitations and Responsible Use

MiniCPM5-1B is a language model that generates content based on learned statistical patterns from training data. It may produce inaccurate, biased, or unsafe outputs, and generated content should be reviewed and verified before use in high-stakes settings.

Users are responsible for evaluating outputs, applying appropriate safeguards, and complying with applicable laws, regulations, and platform policies.

License

This repository and MiniCPM model weights are released under the Apache-2.0 License.

Citation

Please cite our paper if you find our work valuable:

@article{minicpm4,
  title={Minicpm4: Ultra-efficient llms on end devices},
  author={MiniCPM, Team},
  journal={arXiv preprint arXiv:2506.07900},
  year={2025}
}

Abliteration hardware

Recorded in reproduce/reproduce.json:

• GPU: Tesla T4 (14.56 GB VRAM)
• CPU: Intel(R) Xeon(R) CPU @ 2.00GHz
• OS: Linux-6.6.122+-x86_64-with-glibc2.35
• CUDA / driver: CUDA 12.8, driver 580.82.07
• Software: PyTorch 2.11.0+cu128, Heretic 1.4.0

Heretic notes that KL-divergence and refusal numbers can vary slightly with hardware and library versions.