pear
Training Reasoning Models in Simulation to Natively Use Real Robots
pear launched on 13 November 2026. It ranked #116 on Daily Launch on 20 November 2026 with a Launch Score of 14, from +46 stars today on GitHub. It moved down 75 places from the day before.
What is pear?
PEAR is a post-training engine for agentic robotics: it trains reasoning models to natively use robots, with reinforcement learning in simulation. PEAR is the training infrastructure behind Training Reasoning Models in Simulation to Natively Use Real Robots, where an open 27B reasoning model learns to use real robots entirely in simulation, with zero demonstrations.
From its README on GitHub
pear is a repo by mit-han-lab on GitHub. It is written in Python and released under the Apache-2.0 license.
How to get pear
Get it from its repository on GitHub.
Launch timeline
Launch Score, last 7 days
#110 of 226 in the roundup for the week of 16-22 November 2026.