research-log

Frontier Reasoning Agents fail on Interactive 2D Mazes

Frontier Reasoning Agents fail on Interactive 2D Mazes
In: research-log

In collaboration with the team at Fig, we're excited to share an early research preview of MultiNet 2.0.

We put 3 highly capable models in 50 simple 2-dimensional mazes each. Claude Opus 4.8, Kimi K2.6, and Qwen 3.6-27B together solved just 6 out of 150 maze runs, and 45 out of the 50 mazes were not solved by any model.

Evaluation Website & Playable Demo

Full Technical Report

More from Manifold Research
Great! You’ve successfully signed up.
Welcome back! You've successfully signed in.
You've successfully subscribed to Manifold Research.
Your link has expired.
Success! Check your email for magic link to sign-in.
Success! Your billing info has been updated.
Your billing was not updated.