Instructions to use Vita0818/Vireqo-27T-Plus-260818 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use Vita0818/Vireqo-27T-Plus-260818 with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf Vita0818/Vireqo-27T-Plus-260818
# Run inference directly in the terminal:
llama cli -hf Vita0818/Vireqo-27T-Plus-260818

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf Vita0818/Vireqo-27T-Plus-260818
# Run inference directly in the terminal:
llama cli -hf Vita0818/Vireqo-27T-Plus-260818

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf Vita0818/Vireqo-27T-Plus-260818
# Run inference directly in the terminal:
./llama-cli -hf Vita0818/Vireqo-27T-Plus-260818

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf Vita0818/Vireqo-27T-Plus-260818
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Vita0818/Vireqo-27T-Plus-260818

Use Docker

docker model run hf.co/Vita0818/Vireqo-27T-Plus-260818

LM Studio
Jan

vLLM

How to use Vita0818/Vireqo-27T-Plus-260818 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Vita0818/Vireqo-27T-Plus-260818"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Vita0818/Vireqo-27T-Plus-260818",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Vita0818/Vireqo-27T-Plus-260818

Ollama
How to use Vita0818/Vireqo-27T-Plus-260818 with Ollama:
```
ollama run hf.co/Vita0818/Vireqo-27T-Plus-260818
```
Unsloth Desktop

How to use Vita0818/Vireqo-27T-Plus-260818 with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Vita0818/Vireqo-27T-Plus-260818

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "Vita0818/Vireqo-27T-Plus-260818"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use Vita0818/Vireqo-27T-Plus-260818 with Docker Model Runner:
```
docker model run hf.co/Vita0818/Vireqo-27T-Plus-260818
```

Lemonade

How to use Vita0818/Vireqo-27T-Plus-260818 with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Vita0818/Vireqo-27T-Plus-260818

Run and chat with the model

lemonade run user.Vireqo-27T-Plus-260818-{{QUANT_TAG}}

List all available models

lemonade list

Hermes Agent

How to use Vita0818/Vireqo-27T-Plus-260818 with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Vita0818/Vireqo-27T-Plus-260818

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default Vita0818/Vireqo-27T-Plus-260818

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use Vita0818/Vireqo-27T-Plus-260818 with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Vita0818/Vireqo-27T-Plus-260818

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "Vita0818/Vireqo-27T-Plus-260818" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

Vireqo-27T-Plus-260818 / README.md

Vita0818

Release Vireqo-27T-Plus-260818

6254961 verified 9 days ago

preview code

Raw

History Blame Contribute Delete

2.22 kB

	---
	license: apache-2.0
	library_name: llama.cpp
	pipeline_tag: text-generation
	language: [zh, en]
	base_model:
	- Qwen/Qwen3.8-27B
	- prism-ml/Ternary-Bonsai-27B-gguf
	tags: [gguf, reasoning, bounded-thinking, ternary, q2_0, llama-cpp, lm-studio, experimental]
	inference: false
	---

	# Vireqo-27T-Plus-260818

	`Vireqo-27T-Plus-260818` is the ternary bounded-Thinking product built on the unchanged `Vireqo-27B-Plus-260816` payload. It combines the Plus language operating point with the independently validated `Think512-Concise` LM Studio preset.

	中文摘要：这是独立的 T-Plus 产品线，不是把 Thinking 直接塞回原 Plus 模型。原 Plus 权重、SHA 和产品定位保持不变；本目录通过软链接复用主权重。

	## Bundle

	\| Item \| Value \|
	\|---\|---\|
	\| Main file \| `Vireqo-27T-Plus-260818.gguf` \|
	\| Physical source \| unchanged `Vireqo-27B-Plus-260816.gguf` \|
	\| Size \| 7,585,332,288 bytes / 7.0644 GiB \|
	\| SHA-256 \| `a32a8ec286a11c6534bf29d1ee20bd4c02064032b51ae8310bb1216e2de17e03` \|
	\| Preset \| `thinking-preset.json` \|
	\| New physical main-weight copy \| no \|

	The internal `general.name` remains `Vireqo-27B-Plus-260816`; the T-Plus identity is supplied by the release bundle and LM Studio key.

	## Accepted preset

	Think On; Reasoning Budget `512`; maximum response `768`; temperature `0`; top-p `1`; repeat penalty `1.08`; context `2048`; parallel `1`.

	Budget Message:

	```text
	思考预算已到。复核已有结论后，只输出最终答案；禁止重复或展示思考过程。
	```

	## Validation

	Capital, multiplication, and chicken/rabbit final answers were all correct with separated nonempty reasoning and final content. See [`thinking-validation.md`](thinking-validation.md). Plus tends to use more of the 512-token reasoning budget and is slower than standard T.

	## Limitations

	- Unrestricted Thinking remains unsupported.
	- This concise preset is not a general reasoning benchmark claim.
	- Complex tasks may require a future larger-budget line.
	- The underlying ternary model retains all Plus limitations.

	See [`LM-STUDIO-使用指南.md`](LM-STUDIO-使用指南.md), [`TECHNICAL_README.md`](TECHNICAL_README.md), and [`bundle-provenance.json`](bundle-provenance.json).