LLM Schelling Segregation : second mesa-llm example (GSoC 2026)

[abhinavk0220]

Hi everyone!

Following up on my earlier LLM Opinion Dynamics example (#360), I've
just opened a second mesa-llm example — an LLM-powered Schelling
Segregation model (PR #363).

Schelling (1971) is one of the most cited ABM papers ever. The classical
model uses a fixed tolerance threshold to decide if an agent moves. This
version replaces that threshold with LLM reasoning — agents describe their
neighborhood in natural language and decide whether they feel comfortable
staying or want to relocate.

The interesting question is: does the same emergent segregation still
appear when driven by genuine reasoning rather than a fixed number?

Both examples are part of my GSoC 2026 exploration around the
"Mesa Examples Revival" and "Mesa-LLM Iteration" project ideas.

Would love feedback from maintainers and the community!

PRs:

• Mesa 4.x compatibility fix: fix: migrate codebase to Mesa 4.x API (resolves mesa.space removal) mesa-llm#153
• LLM Opinion Dynamics: feat: add LLM Opinion Dynamics example using mesa-llm #360
• LLM Schelling Segregation: feat: add LLM Schelling Segregation example using mesa-llm #363