Where questions become systems.
AXIVELLE Labs is the research division. It studies how models reason, how they are trained and evaluated, and how new capabilities can be measured honestly before they are built into products.
AI research and development.
Scientific, experimental and technical.
Experimental models
Exploring architectures and training methods at small scale first.
Training and evaluation
Building evaluation that reflects real use, not only benchmarks.
Reasoning research
How models plan, verify and correct their own work.
Multimodal research
Shared representations across language, vision and structured data.
Intelligent systems
Models as components of larger, dependable systems.
Emerging technologies
Tracking and testing new ideas as the field moves.
AXIVELLE Labs Workspace
A research workspace for experiments, evaluations, datasets and model comparison.
What AXIVELLE LABS is working on.
Evaluation suites
Reusable evaluations for reasoning, reliability and multimodal tasks.
A future project. Not yet designed in detail.
Research notes
Technical notes describing methods, results and limitations.
A future project. Not yet designed in detail.
Every division mark keeps the AXIVELLE shell — the two legs of the apex — and changes only the core. The shell is the company; the core is the discipline.

