Instructions to use Tomkiel/umba-5.6.600 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use Tomkiel/umba-5.6.600 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 Tomkiel/umba-5.6.600
# Run inference directly in the terminal:
llama cli -hf Tomkiel/umba-5.6.600

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf Tomkiel/umba-5.6.600
# Run inference directly in the terminal:
llama cli -hf Tomkiel/umba-5.6.600

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 Tomkiel/umba-5.6.600
# Run inference directly in the terminal:
./llama-cli -hf Tomkiel/umba-5.6.600

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 Tomkiel/umba-5.6.600
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Tomkiel/umba-5.6.600

Use Docker

docker model run hf.co/Tomkiel/umba-5.6.600

LM Studio
Jan
Ollama
How to use Tomkiel/umba-5.6.600 with Ollama:
```
ollama run hf.co/Tomkiel/umba-5.6.600
```
Unsloth Desktop

How to use Tomkiel/umba-5.6.600 with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Tomkiel/umba-5.6.600

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": "Tomkiel/umba-5.6.600"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use Tomkiel/umba-5.6.600 with Docker Model Runner:
```
docker model run hf.co/Tomkiel/umba-5.6.600
```

Lemonade

How to use Tomkiel/umba-5.6.600 with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Tomkiel/umba-5.6.600

Run and chat with the model

lemonade run user.umba-5.6.600-{{QUANT_TAG}}

List all available models

lemonade list

Hermes Agent

How to use Tomkiel/umba-5.6.600 with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Tomkiel/umba-5.6.600

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 Tomkiel/umba-5.6.600

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use Tomkiel/umba-5.6.600 with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Tomkiel/umba-5.6.600

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 "Tomkiel/umba-5.6.600" \
  --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"

umba-5.6.600 (KAT-Coder-V2.5-Dev 35B MoE)

umba-5.6.600 is a state-of-the-art agentic software engineering model fine-tuned on top of Kwaipilot/KAT-Coder-V2.5-Dev (Sparse MoE 35B total / 3B active parameters).

🧬 Training Architecture

Base Model: Kwaipilot/KAT-Coder-V2.5-Dev (Top Open-Source model on SWE-bench Verified: 69.4%)
Phase 1 (SFT): 600 steps trained on ~~49,500 curated trajectories (~~186k canonical tool calls, strict prompt assumptions, Fable Method, <think> deep CoT).
Phase 2 (DPO): 100 steps of contrastive direct preference optimization to eradicate hallucinated tool calls and language attractors.

📦 APEX LocalAI Quantizations Included

Quantized using LocalAI APEX (Adaptive Precision for EXpert Models):

Filename	APEX Flavor	Target Profile	Best Suited For
`umba-5.6.600-tripod.gguf`	Quality	Q6_K / Q5_K adaptive	Maximum accuracy and coding precision (Production / GPUs with 24GB+ VRAM)
`umba-5.6.600-dirt.gguf`	Compact-IQ4	Q4_K edge / IQ4_XS core	Superior balance between memory usage and code syntax retention
`umba-5.6.600-facelift.gguf`	Apex Compact	Q4_K edge / Q3_K core	High throughput, minimal memory footprint (16GB GPUs / Apple Silicon)

🛠️ Usage with llama-server

llama-server -m umba-5.6.600-tripod.gguf -c 32768 -ngl 99 --port 8080

Downloads last month: 738

GGUF

Model size

35B params

Architecture

qwen35moe

Hardware compatibility

We're not able to determine the quantization variants.

View all variants

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for Tomkiel/umba-5.6.600

Base model

Kwaipilot/KAT-Coder-V2.5-Dev

Quantized

(70)

this model