Vertex AGI is a small, independent AI lab building focused, efficient language models — designed, trained, evaluated, and shipped end-to-end on consumer hardware rather than large-scale compute clusters.
We believe useful, high-quality models don't have to start from massive budgets. Our work spans general-purpose assistants, agentic/tool-using models, and creative and character-driven writing models, built through careful data curation, distillation from stronger teacher models, and fine-tuning.
Every model we publish ships with both MLX and GGUF formats in a single repo, so it runs equally well on Apple Silicon or anywhere the llama.cpp ecosystem (Ollama, LM Studio, and others) reaches.