SOVEREIGNTY STARTS WITH MEASUREMENT
Measurement, diagnostic, and steering instrument designed to detect changes in AI behavior before they become an operational risk.
Privacy-first. No infrastructure changes required. Deployable in sovereign environments. Designed to strengthen the traceability, auditability, and governance of AI systems.
NeoMundi is advancing AI observability toward behavioral metrology : measuring responses under defined conditions, establishing a reference baseline, then detecting when behavior changes.
NeoMundi Research develops an independent observatory to measure the stability, variations, and regime changes of generative AI systems.
NeoMundi Research AI Observatory
The NeoMundi Observatory is an independent observatory with a global ambition. It measures the behavior of AI systems over time to establish a reference baseline and detect variations, drifts, or regime changes.
Monthly Mapping — June 2026 :
A first comparative snapshot of observed behavior across 12 large language models.
In AI, price, stability, and accuracy do not always move together.
Read the article.
Map of 12 systems by observable stability and evaluated factual validity; point color indicates relative cost (green low, orange intermediate, red high).
Observable stability →
Color = relative cost per request
Stability and validity barely predict one another (correlation ≈ 0.32), and cost overlaps with neither. Exploratory map: observed profiles, no public ranking of providers. It does not designate a “best model”; it shows different trade-offs.
NeoMundi · instrument view · June 2026
The AI Oservatory’s latest publications
- July 2026 AI Cartography: Measuring the Behavior of 12 LLMs
- The framework for measuring AIs already exists. It still have to be made operational.
- AI Barometer #4 – Week of July 6-12, 2026
- The AI risk seen from underwriting: why the insurance market is flying blind
- NeoMundi Barometer #3 – Week of June 30 to July 6, 2026
OPEN SCIENCE
We built the instrument.
Now we are launching the observatory.
NeoMundi Research is launching a collaborative program for the continuous observation of generative AI systems.
Its objective: to measure, over time, the stability, drift, errors, costs, and behavioral signatures of major AI systems, in order to make their operation more understandable, more traceable, and more governable
The program will rely in particular on a monthly cartography and a weekly barometer, published according to documented and reproducible protocols.
The ultimate goal: to make AI systems more understandable and more governable, and to build trust based on evidence
To explore the instrument, follow the publications and contribute to the work :
The NeoMundi Instrument / ControlTower
→ Discover the real-time measurement and monitoring instrument for AI behaviour
Scientific Program
→ https://github.com/neomundi-io/neomundi-ai-observatory
Industrial Specifications
→ github.com/neomundi-io
ORIENTATION COMMITTEE IN FORMATION
NeoMundi Research is progressively bringing together complementary profiles to challenge the methodology, measurement protocols and published results.
Objective: to strengthen the scientific, technical and institutional rigor of the generative AI observation program.
First Members:
- Joël Ignasse, Scientific journalist for La Recherche. Author of “Sur les traces du Nouveau T.rex” (Eyrolles, 2025).
- Cédric Chatelain, Quality consultant and IRCA-certified auditor, specialized in audit, validation, and ISO standards. Based in Bern, Switzerland.
- James Aull, Founder of ASRO™, an independent governed-state witness and evidence layer for AI systems. Based in Twin Lake, Michigan, United States.
- Inès Ramoul, GDPR & AI Act legal specialist, focused on digital law, data protection and AI governance. Legal and compliance contributor to the NeoMundi AI Observatory. Based in Lyon, France.