Instructions to use genomalabs/kalypso-v1.2l with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use genomalabs/kalypso-v1.2l 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 genomalabs/kalypso-v1.2l:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf genomalabs/kalypso-v1.2l:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf genomalabs/kalypso-v1.2l:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf genomalabs/kalypso-v1.2l:Q4_K_M

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 genomalabs/kalypso-v1.2l:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf genomalabs/kalypso-v1.2l:Q4_K_M

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 genomalabs/kalypso-v1.2l:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf genomalabs/kalypso-v1.2l:Q4_K_M

Use Docker

docker model run hf.co/genomalabs/kalypso-v1.2l:Q4_K_M

LM Studio
Jan

vLLM

How to use genomalabs/kalypso-v1.2l with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "genomalabs/kalypso-v1.2l"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "genomalabs/kalypso-v1.2l",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/genomalabs/kalypso-v1.2l:Q4_K_M

Ollama
How to use genomalabs/kalypso-v1.2l with Ollama:
```
ollama run hf.co/genomalabs/kalypso-v1.2l:Q4_K_M
```

Unsloth Studio

How to use genomalabs/kalypso-v1.2l 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 genomalabs/kalypso-v1.2l 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 genomalabs/kalypso-v1.2l to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for genomalabs/kalypso-v1.2l to start chatting

How to use genomalabs/kalypso-v1.2l with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf genomalabs/kalypso-v1.2l:Q4_K_M

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": "genomalabs/kalypso-v1.2l:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use genomalabs/kalypso-v1.2l with Docker Model Runner:
```
docker model run hf.co/genomalabs/kalypso-v1.2l:Q4_K_M
```

Lemonade

How to use genomalabs/kalypso-v1.2l with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull genomalabs/kalypso-v1.2l:Q4_K_M

Run and chat with the model

lemonade run user.kalypso-v1.2l-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use genomalabs/kalypso-v1.2l with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf genomalabs/kalypso-v1.2l:Q4_K_M

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 genomalabs/kalypso-v1.2l:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use genomalabs/kalypso-v1.2l with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf genomalabs/kalypso-v1.2l:Q4_K_M

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 "genomalabs/kalypso-v1.2l:Q4_K_M" \
  --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"

KALYPSO v1.2L

GENOMA Labs' public coding + planning model. Qwen3-Coder-30B-A3B-Instruct (30.5B total / 3.3B active MoE) with a GENOMA failure-handling LoRA (r16 on attention + expert MLPs). Successor to kalypso-v1.1l.

Why upgrade from v1.1L

Identical harness, identical budgets, all rows computed (never narrated):

Benchmark (GENOMA harness)	v1.1L (14B dense)	v1.2L (30B-A3B)
Coding — v4_hard pass@1, hidden-test sandbox (41 tasks)	57.7% (N=3)	73.2% (N=2)
Plan-grade orchestration (30 tasks, calibrated LLM-judge)	0.832	0.848
Execution-grade orchestration (failure-handling outcome judge)	0.437	0.482
MMLU-Pro (n=300)	41.0%	50.3%
Active parameters per token	14B	3.3B (~4× cheaper inference)

Better on every axis at a fraction of the serving cost. 32k context (base-native), Apache-2.0.

What the fine-tune is (and honest limits)

Corpus (800 examples, fully public-safe): 320 failure-handling dialogues + 160 orchestration plans with explicit verification gates and rollback/compensation steps + 320 coding-hold examples. All teachers generic and local; zero proprietary content (audited).
Honest note: the uplift claim in the table above is vs KALYPSO v1.1L, not vs the raw Qwen base. Against its own base under identical settings, this fine-tune measures −7pp on coding pass@1 (base 82.9% vs 73.2%, N=2 each) and parity on planning/execution/knowledge. If raw one-shot coding is your only criterion, the unmodified base is the stronger pick; v1.2L remains far ahead of v1.1L on every axis. We publish all numbers as measured.
Execution-grade planning is an open weakness industry-wide — every model we measured (ours included, plus our previous public model) scores < 0.5 on "does the plan actually handle the failure mode (rollback / compensation / gate-before-prod)". We release our benchmark methodology so the community can push on this.

Eval methodology

Coding: the model must emit implementation + tests as fenced files; hidden pytest tests are injected and run in a sandbox; pass = tests green. Single-shot (pass@1), temperature 0.
Plan-grade: 30 orchestration scenarios; a calibrated LLM-judge scores semantic plan quality (judge discrimination margin verified before use).
Execution-grade: an outcome judge mentally executes each plan and scores goal-achievement (0–60) + failure-mode handling (0–40); calibrated to penalize happy-path-only plans. All models graded by the same judge.
N-run where stated; single-run rows are marked by their N.

Serving

Fits a single 24 GB GPU in 4-bit (~19 GB); vLLM/llama.cpp/ollama supported (GGUF provided).
Recommended: temperature 0–0.2 for coding/planning.

Provenance & license

Base: Qwen/Qwen3-Coder-30B-A3B-Instruct (Apache-2.0). Fine-tune: GENOMA Labs, Apache-2.0. No robotics, CAN-bus, or proprietary GENOMA data in the training corpus (audited pre-release).

GENOMA Labs — 2026-08.

Downloads last month: -

GGUF

Model size

31B params

Architecture

qwen3moe

Hardware compatibility

4-bit

Model tree for genomalabs/kalypso-v1.2l

Base model

Qwen/Qwen3-Coder-30B-A3B-Instruct

Quantized

(164)

this model