Instructions to use gsstec/G1-360M_V18.2_PRODUCTION_F16 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use gsstec/G1-360M_V18.2_PRODUCTION_F16 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 gsstec/G1-360M_V18.2_PRODUCTION_F16:F16
# Run inference directly in the terminal:
llama cli -hf gsstec/G1-360M_V18.2_PRODUCTION_F16:F16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf gsstec/G1-360M_V18.2_PRODUCTION_F16:F16
# Run inference directly in the terminal:
llama cli -hf gsstec/G1-360M_V18.2_PRODUCTION_F16:F16

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 gsstec/G1-360M_V18.2_PRODUCTION_F16:F16
# Run inference directly in the terminal:
./llama-cli -hf gsstec/G1-360M_V18.2_PRODUCTION_F16:F16

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 gsstec/G1-360M_V18.2_PRODUCTION_F16:F16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf gsstec/G1-360M_V18.2_PRODUCTION_F16:F16

Use Docker

docker model run hf.co/gsstec/G1-360M_V18.2_PRODUCTION_F16:F16

LM Studio
Jan

vLLM

How to use gsstec/G1-360M_V18.2_PRODUCTION_F16 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "gsstec/G1-360M_V18.2_PRODUCTION_F16"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "gsstec/G1-360M_V18.2_PRODUCTION_F16",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/gsstec/G1-360M_V18.2_PRODUCTION_F16:F16

Ollama
How to use gsstec/G1-360M_V18.2_PRODUCTION_F16 with Ollama:
```
ollama run hf.co/gsstec/G1-360M_V18.2_PRODUCTION_F16:F16
```

Unsloth Studio

How to use gsstec/G1-360M_V18.2_PRODUCTION_F16 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 gsstec/G1-360M_V18.2_PRODUCTION_F16 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 gsstec/G1-360M_V18.2_PRODUCTION_F16 to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for gsstec/G1-360M_V18.2_PRODUCTION_F16 to start chatting

Docker Model Runner
How to use gsstec/G1-360M_V18.2_PRODUCTION_F16 with Docker Model Runner:
```
docker model run hf.co/gsstec/G1-360M_V18.2_PRODUCTION_F16:F16
```

Lemonade

How to use gsstec/G1-360M_V18.2_PRODUCTION_F16 with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull gsstec/G1-360M_V18.2_PRODUCTION_F16:F16

Run and chat with the model

lemonade run user.G1-360M_V18.2_PRODUCTION_F16-F16

List all available models

lemonade list

Atomic Chat

G1-360M V18.2 — Production F16

G1-360M V18.2 is a production-oriented enterprise AI model developed by Gaston Software Solutions LLP (GSS-LLP) for practical business, enterprise software, workflow, API, data, security, and architecture reasoning.

G1 is designed to operate as an enterprise AI assistant that helps users reason about business systems and operational processes while maintaining explicit assumptions and avoiding unsupported claims.

Runs on 3 GB RAM. G1-360M V18.2 (GGUF F16) can be loaded and run on devices with as little as 3 GB of available RAM, making it practical for local workstations, laptops, edge servers, and offline enterprise environments without requiring dedicated GPU hardware.

Model journey

Developer

Gaston Software Solutions LLP (GSS-LLP)

Registration No.: 80041130611335
Location: Kampala, Uganda
Website: https://www.gss-tec.com
Email: info@gss-tec.com
Telephone: +256 755 274955
LinkedIn: Gaston Software Solutions LLP

Model identity

Property	Value
Model	G1-360M V18.2
Release	Production F16
Developer	Gaston Software Solutions LLP
Model class	Compact enterprise language model
Parameters	~360M
Format	GGUF
Precision	F16
Primary task	Text generation / enterprise reasoning
Deployment	Local, private, enterprise and application-integrated inference
License	G1-360M Commercial License

What is G1?

G1 is an enterprise-focused language model built for practical reasoning over business and software-system problems.

Its intended domain includes:

Enterprise software architecture
Business workflows
API and integration reasoning
Data validation
Security controls and authorization reasoning
Operational processes
Business rules and constraints
Decision reasoning
Enterprise systems
Structured problem solving
Offline and resilient workflows
Data and identifier validation
Application and system behavior
Business-process analysis

G1 is intended to provide concise, practical answers that can be incorporated into enterprise workflows, internal applications, software tools, and AI-assisted business processes.

Enterprise capabilities

G1 can be used to assist with:

Business systems

Reason about business processes, transaction flows, operational rules, identifiers, records, and system behavior.

Enterprise architecture

Analyze components, lifecycle states, interfaces, dependencies, workflows, validation boundaries, and system behavior.

APIs and integrations

Reason about API inputs, outputs, validation, authorization, integration failures, workflow sequencing, and service dependencies.

Security reasoning

Identify practical controls such as input validation, authorization, access boundaries, and protection of sensitive operations.

G1 should not be treated as a replacement for professional security assessment or legal/compliance review.

Data validation

Reason about required fields, known identifiers, valid values, relationships, and acceptance rules before data enters an enterprise workflow.

Decision reasoning

Compare practical alternatives, identify priorities, explain trade-offs, and select actions based on stated business constraints.

Constraint following

Follow explicit numerical, structural, procedural, and formatting constraints when they are provided in the prompt.

Offline resilience

Reason about business continuity when external APIs, networks, or other dependencies become unavailable.

Enterprise operations

Help structure practical operational responses, validation procedures, workflow controls, and recovery processes.

Intended use

G1-360M V18.2 is intended for lawful use by:

Enterprises
Corporations
Government and institutional organizations
Software companies
Developers
System integrators
Internal AI teams
Research and development teams
Business-process teams
Enterprise architecture teams

Organizations may integrate G1 into internal applications, enterprise software, APIs, local AI systems, automation workflows, and customer-facing products subject to the applicable license.

Fine-tuning and customization

The G1 license permits organizations to adapt the model for their own business requirements, including:

Fine-tuning
Domain adaptation
Instruction tuning
Enterprise-specific customization
Internal evaluation
Application integration
Workflow-specific optimization
Development of specialized derivatives

Organizations can therefore adapt G1 for their own domain while retaining the rights and responsibilities defined by the G1 license.

Deployment

This release is distributed as a GGUF F16 model and is intended for local inference environments supporting GGUF models.

Typical deployment environments include:

llama.cpp
Ollama
Local AI runtimes supporting GGUF
Enterprise inference servers
Private infrastructure
Development workstations
Controlled offline environments

Example use cases

G1 can be integrated into applications such as:

Enterprise help assistants
Internal business copilots
Workflow assistants
API troubleshooting assistants
Enterprise architecture assistants
Data-quality assistants
Operational support tools
Business-process assistants
Internal knowledge applications
Offline enterprise AI systems
Software development assistants
Customer-service workflow tools

Example interaction

User:

A service receives an identifier from an untrusted caller. What should happen before the identifier is used to access protected data?

G1:

Validate that the identifier is valid and authorized for the requested operation before using it to access protected data.

The model is intended to prioritize practical controls and explicit reasoning over unsupported implementation assumptions.

Behavioral design

G1 is designed to emphasize:

Instruction following
Constraint adherence
Arithmetic correctness
Self-correction
Offline resilience
Decision reasoning
Enterprise architecture reasoning
Security reasoning
Data validation

These behaviors are intended to improve reliability in practical enterprise workflows.

Limitations

G1-360M is a compact language model. It can produce incorrect, incomplete, outdated, or overly simplified answers.

Users should independently verify outputs when they affect:

Financial decisions
Legal decisions
Security decisions
Regulatory compliance
Safety-critical systems
Production infrastructure
Medical decisions
High-impact organizational decisions

G1 does not have inherent access to private organizational systems, databases, APIs, or current external information unless an application explicitly provides such access.

The model should not be assumed to know facts that were not provided in its context.

Security considerations

G1 should be deployed with appropriate application-level security controls.

In particular:

Do not treat model output as authorization.
Validate model-generated actions before execution.
Apply authentication and authorization outside the model.
Restrict tool and API permissions.
Validate structured outputs before execution.
Do not expose secrets through prompts or model context.
Log and monitor production integrations appropriately.

The model itself should not be used as the sole security boundary.

Training and development

G1-360M V18.2 is part of the G1 model development series developed by GSS-LLP.

The production release incorporates iterative model development and behavioral-repair training intended to improve practical enterprise behavior, including constraint following, arithmetic, validation, security reasoning, decision reasoning, enterprise architecture, offline resilience, instruction following, and self-correction.

License

G1-360M V18.2 is distributed under the G1-360M Commercial License.

The complete license is available at:

https://www.g1-license.gss-tec.com/

The license permits corporate and enterprise use and permits organizations to fine-tune and customize the model for their business requirements, subject to the terms of the license.

License: G1-360M Commercial License

Attribution

When redistributing the original model or derivative distributions where attribution is required by the license, retain attribution to:

Gaston Software Solutions LLP (GSS-LLP) Registration No. 80041130611335 Kampala, Uganda

Contact

Gaston Software Solutions LLP (GSS-LLP)

Website: https://www.gss-tec.com Email: info@gss-tec.com Telephone: +256 755 274955 LinkedIn: Gaston Software Solutions LLP

Disclaimer

G1 is provided for general information, development, research, and enterprise software use. Model outputs should be reviewed and validated before being used in consequential systems or decisions.

GSS-LLP makes no representation that every generated response is accurate, complete, or suitable for a particular purpose.

G1-360M V18.2 Developed by Gaston Software Solutions LLP (GSS-LLP) Kampala, Uganda

Downloads last month: 5

GGUF

Model size

0.4B params

Architecture

llama

Hardware compatibility

16-bit