Let’s build together We build in the open. We share models and ideas because intelligence belongs to everyone.
Every leap forward starts with a different way of seeing. At Aixon Lab, we believe that progress should be shared. Our open-source approach gives more people the tools to explore, learn, and shape artificial intelligence for themselves.
Through advanced fine-tuning, custom LoRA development, and innovative workflows, we make AI more accessible, more expressive, and more effective for the people building with it.
The right capabilities for the task at hand.
Open models, thoughtfully refined.
Technology in service of imagination.
Ideas don’t exist in isolation. They meet, evolve, and become something greater. Step inside our interpretation of a connected intelligence.
MANY CONNECTIONS.
ONE EVOLVING INTELLIGENCE.
We shape the model around your ambition.
Explore three ways to build with Aixon.
Language models that speak your domain. Built on foundations including Mistral, Llama, and NVIDIA NeMo, our fine-tuned LLMs bring specialized capabilities to technical documentation, customer support, and analysis.
Custom response patterns. Deeper domain knowledge. A thoughtful balance of performance and efficiency.
Explore language modelsYour visual language, made possible. Our custom LoRAs for FLUX and Stable Diffusion XL capture distinctive styles and capabilities in compact, efficient model adaptations.
From the warm, dreamy grading of contemporary cinema to the textured character of 1940s film.
Explore visual modelsConnect the possibilities. We develop ComfyUI workflows that bring models and creative processes together, making sophisticated AI techniques easier to put into practice.
Purposeful pipelines. Repeatable processes. More room to create.
Discuss a workflowA collection of fine-tuned language models
and expressive visual adaptations.
6 models found
Aether-12b is a fine-tuned large language model based on Arcanum-12b, further trained on the CleverBoi-Data-20k dataset.
Explore model
Zinakha 12b tries to become the perfect companion for any chat which involves multiple roles. The ability to understand context is pretty awesome and excels in creativity and storytelling. It is built on Nemo 12b and trained on different datasets as well as some layer merges to enhance its capabilities.
Explore model
Grey-12b is a merged language model created by combining multiple models using the della_linear merge method, with Aether-12b as the base model.
Explore model
Zara 14b tries to become the perfect companion for any chat which involves multiple roles. The ability to understand context is pretty awesome and excels in creativity and storytelling. It is built on Lamarck 14B v0.7 and trained on different datasets as well as some layer merges to enhance its capabilities.
Explore model
VISUAL INTELLIGENCEA / 05Cinematic-Octane is a LoRA model meticulously fine-tuned to embody the warm, dreamy color grading prevalent in contemporary cinema.
Explore model
VISUAL INTELLIGENCEA / 06Cinematic-1940s is a LoRA model finetuned to capture the essence of 1940s cinema, specifically designed to enhance the vintage film texture.
Explore modelGOOD IDEAS GO FURTHER WHEN THEY’RE SHARED.
Have a model in mind, a workflow to rethink,
or a possibility worth exploring? We’re listening.
