Prague, November 16th, 2025 - Google DeepMind just released SIMA 2 and we’re excited to share that Space Engineers took part in their research. Thanks to the DeepMind team for the opportunity!
The first version of SIMA learned to perform over 600 language-following skills, such as "turn left", "climb the ladder", and "open the map", across a diverse set of commercial video games. It operated in these environments as a human player would - by looking at the screen and using a virtual keyboard and mouse to navigate, without any access to internal game mechanics.
With SIMA 2, the DeepMind team moved beyond instruction-following. By embedding a Gemini model as the agent's core, SIMA 2 can do more than simply respond to instructions - it can think and reason about them.
Here’s SIMA 2 exploring and interacting inside the complex voxel environment of Space Engineers.
Space Engineers poses a unique challenge for AI agents, because it features a realistic, volumetric-based physics engine. Everything in the game can be assembled, disassembled, damaged, and destroyed, creating a dynamic environment that requires genuine problem solving and spatial reasoning.
- https://deepmind.google/blog/sima-2-an-agent-that-plays-reasons-and-learns-with-you-in-virtual-3d-worlds/
- https://www.spaceengineersgame.com/
🛠️ Feedback & Support - https://support.keenswh.com/
🛰️ Discord: https://discord.gg/keenswh
📫 Newsletter: https://www.keenswh.com/newsletter
🚀 Community Hub: https://www.spaceengineersgame.com/community-hub/
➡️ Connect and find us on your platform of choice: https://www.keenswh.com/connect/