# A generalist AI agent for 3D virtual environments

*Source* · `src-deepmind-2024-sima-blog` · status: **n/a**

| field | value |
|---|---|
| kind | Technical report / blog |
| authors | Google DeepMind SIMA Team |
| venue | Google DeepMind blog |
| date | 2024-03-13 |
| url | https://deepmind.google/blog/sima-generalist-ai-agent-for-3d-virtual-environments/ |
| accessed | 2026-10-24 |

## Notes

600 basic skills; nine commercial games (Valheim, Teardown, No Man's Sky, Wobbly Life, Satisfactory, Goat Simulator 3, Hydroneer, Space Engineers, Eco) plus four research environments. Generalist outperformed per-game specialists; held-out-game agent nearly matched specialist. Tasks about 10 seconds; no multi-step strategic planning.

## Backlinks (computed)

**cited by:** [Generalist instructable game agents (SIMA)](../technique/generalist-game-agents.md)

## Citation

Google DeepMind SIMA Team (2024-03-13). *A generalist AI agent for 3D virtual environments*. Google DeepMind blog https://deepmind.google/blog/sima-generalist-ai-agent-for-3d-virtual-environments/

---
Provenance: created in run `2026-10-24-techniques-llm-agents`.