Breaking

OpenAI Shows Astrophysicist Using Codex to Build Black Hole Simulations

June 10, 2026 at 18:13 EDT

AI Agents
Software Dev & Coding
Research & Papers

On May 29, 2026, OpenAI published a case study on its official blog describing how computational astrophysicist Chi-kwan Chan uses the code-generation AI "Codex" to build black hole simulations. The work is presented as support for studying extreme physical phenomena and testing Einstein's theory of General Relativity.

AI in Science · OpenAI Codex

Coding the Cosmos: AI Helps Simulate Black Holes

Astrophysicist Chi-kwan (CK) Chan of the University of Arizona is using OpenAI's code-generation model Codex to build black-hole simulations — work aimed at testing Einstein's theory of general relativity.

~25

researchers at a Codex hackathon for astronomy, physics & applied math

real-problem projects built and shared with Codex

Petabytes

of telescope data behind the simulations Chan develops

From observation to a relativity test — the workflow

Event Horizon Telescope

Global radio array images a black hole's shadow

→

Codex-assisted code

AI helps write specialized GPU simulation code

→

Test general relativity

Compare models of Sgr A* against Einstein's theory

What makes this notable

Codex is applied not as a general coding aid, but to the specialized numerical computation of frontier astrophysics.

Valued like code completion and knowledge retrieval for researchers
Accuracy improves when prompted to reference the actual equations
A real-world example of embedding AI code generation into basic science

Note: The showcase presents no specific benchmarks or performance figures — it stands as an illustrative, not measured, demonstration.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.