Central banks and policy teams
Comparing monetary-rule coefficients and welfare costs on models whose determinacy has been checked, not assumed.
Industry Intelligence · Macroeconomics
GenXEcon
AI-driven DSGE modelling and macroeconomic analysis.
GenXEcon is an AI-driven platform for DSGE modelling and macroeconomic analysis that turns a plain-language description of an economy into a solved, mathematically verified model. Describe an economy in plain language and get back a Dynamic Stochastic General Equilibrium model that actually solves. Four specialised agents design the structure and write the equations; a symbolic compiler then solves the system and checks determinacy, and anything that fails is rejected rather than returned with a caveat. The house phrase for it is that the language model proposes and verified mathematics disposes.
Who it is for
Built for a specific job
Academic faculties
A shared team model library, SSO and role-based access for a department running many projects at once.
Applied economists and students
A free tier with local models and your own API key, and Python export so nothing you build is trapped in the app.
What it does
Capabilities
Verified, not just generated
Every model is compiled symbolically and checked for determinacy. The pipeline self-repairs, and rejects what fails.
Impulse responses and moments
Simulation with parameter overrides, plus exact business-cycle properties — volatility, persistence, comovement.
Policy and welfare analysis
Compare monetary-rule coefficients across specifications and compute Lucas-style welfare costs.
Sixteen econometric methods
HP and Baxter-King filters, VAR, SVAR and cointegration tests, in one toolbox alongside the modelling.
Estimation that is honest about identification
Maximum likelihood and Bayesian MCMC. Cholesky orderings are required rather than silently defaulted; sign restrictions are labelled as set identification.
Leave whenever you want
Publication-ready PDF reports, and standalone Python code that runs with no dependency on the platform.
How it works
End to end
The theory economist designs
An agent turns the description of the economy into a model structure — agents, frictions, shocks.
The computational economist writes it
A second agent produces equations that can actually be solved, rather than a plausible-looking system.
The compiler solves and verifies
Symbolic log-linearisation, Klein's QZ method, and a Blanchard-Kahn determinacy check. Failure here ends the run.
The reviewer validates
A fourth agent checks the impulse responses against what economics says they should look like before anything is returned.
Under the hood
Stack and standards
- Klein (2000) QZ method
- Blanchard-Kahn determinacy
- Metropolis-Hastings MCMC
- Kalman filter likelihood
- Symbolic log-linearisation
- 279 offline tests
- Five verified reference models
What it will not do
GenXEcon publishes its constraints instead of leaving you to discover them. Solutions are first order only, which means no risk adjustment. Recursive SVAR identification carries the limitations recursive SVARs always carry. Verification is stated concretely — agreement to 1e-6 against analytical solutions — rather than as a claim of general correctness.
Pricing
What it costs
Researcher
FreeStudents and individual researchers
- Local models
- Bring your own LLM API key
- Community support
- Full analysis toolbox
Professional
$49per month
- Hosted workspace
- Managed generation
- Priority compute
- Email support
Institution
CustomCentral banks, faculties, research teams
- Shared team library
- SSO
- Dedicated engineer
Academic and non-profit discounts are available, and every model exports as standalone Python so nothing you build is trapped in the platform.
Questions
Frequently asked questions
What is GenXEcon?
GenXEcon turns a plain-language description of an economy into a solved Dynamic Stochastic General Equilibrium (DSGE) model. Four specialised agents design the structure and write the equations, and a symbolic compiler then solves and verifies the system.
Does GenXEcon verify the models it generates?
Yes. Every model is compiled symbolically, solved using Klein's (2000) QZ method, and checked for determinacy against the Blanchard-Kahn conditions. Models that fail are rejected rather than returned with a caveat.
How much does GenXEcon cost?
The Researcher tier is free, using local models and your own LLM API key. Professional is $49 per month. Institution pricing is custom. Academic and non-profit discounts are available, and every model exports as standalone Python so nothing you build is trapped in the platform.
Can I export a model out of GenXEcon?
Yes. GenXEcon produces publication-ready PDF reports and standalone Python code that runs with no dependency on the platform.
What are GenXEcon's stated limitations?
Solutions are first order only, which means no risk adjustment. Recursive SVAR identification carries the limitations recursive SVARs always carry. Verification is stated concretely as agreement to 1e-6 against analytical solutions rather than as a claim of general correctness.
What econometric methods are included?
Sixteen, including HP and Baxter-King filters, VAR, SVAR and cointegration tests, plus maximum likelihood and Bayesian MCMC estimation with Kalman-filter likelihoods.
Also from Genexsus
Related products
Industry Intelligence
GenXFintel
The transparent AI analyst for finance teams.
Industry Intelligence
GenXChains
AI-powered sales and operations planning.
AI Platform
GenXAI
A graph-native agentic AI framework.
See GenXEcon on your own data
A demo with your numbers in it tells you more than any page can. If it turns out GenXEcon is not the right fit, we will tell you that too.