Instructions to use jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus 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 jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus
# Run inference directly in the terminal:
llama cli -hf jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus
# Run inference directly in the terminal:
llama cli -hf jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus

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 jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus
# Run inference directly in the terminal:
./llama-cli -hf jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus

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 jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus
# Run inference directly in the terminal:
./build/bin/llama-cli -hf jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus

Use Docker

docker model run hf.co/jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus

LM Studio
Jan

vLLM

How to use jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus

Ollama
How to use jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus with Ollama:
```
ollama run hf.co/jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus
```

Unsloth Studio

How to use jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus to start chatting

How to use jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus

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 jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus

Run Hermes

hermes

OpenClaw new

How to use jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus

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 "jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus" \
  --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"

Docker Model Runner
How to use jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus with Docker Model Runner:
```
docker model run hf.co/jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus
```

Lemonade

How to use jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull jcbtc/Step-3.7-Flash-ROCmFPX-Q3-QualityPlus

Run and chat with the model

lemonade run user.Step-3.7-Flash-ROCmFPX-Q3-QualityPlus-{{QUANT_TAG}}

List all available models

lemonade list

Atomic Chat

Step-3.7-Flash-ROCmFPX-Q3-QualityPlus

File size: 1,273 Bytes

{
  "model": "Step-3.7-Flash-ROCmFPX-Q3-QualityPlus",
  "benchmark": "Tool-Eval full",
  "harness_version": "2.0.7",
  "run_id": "2026-07-05T06-07-28.330900Z_405f5a32",
  "scenario_count": 69,
  "final_score": 88,
  "deployability": 70,
  "responsiveness": 27,
  "total_points": 122,
  "max_points": 138,
  "status_counts": {
    "pass": 57,
    "partial": 8,
    "fail": 4
  },
  "runtime": {
    "backend": "llama.cpp ROCmFPX / Vulkan",
    "context": 65536,
    "mtp": true,
    "speculative_n_max": 2,
    "speculative_p_min": 0.75,
    "target_kv": "q8_0/q8_0",
    "draft_kv": "q8_0/q8_0",
    "batch": 8192,
    "ubatch": 2048
  },
  "speed_and_memory": {
    "pp_tps": 145.65656394243823,
    "tg_tps": 32.338072639262656,
    "prompt_tokens": 34333,
    "predicted_tokens": 33430,
    "peak_pooled_gpu_gib": 96.34914779663086,
    "peak_ram_used_gib": 107.00336074829102,
    "wall_time": "21:21.23"
  },
  "notes": [
    "Local AMD Ryzen AI Max+ 395 / Strix Halo measurement.",
    "Uses the Step native tool_response chat template with protocol-boundary escaping.",
    "The public llm.ciru.ai StepFun tool-eval page currently documents the Step tool-calling methodology and matching FP4 score row; this JSON records the exact Q3 QualityPlus run summary."
  ]
}