Category

Spatial Reasoning: RL environments and AI benchmarks

AI benchmarks
30
Shares entries with
0 categories

Snapshot 2026-10-11. How categories are built · Where the data comes from

In short

What are RL environments for spatial reasoning?
A simulated workplace where an AI agent practises a task (a code repository, a browser, business software) and an automatic grader scores each attempt. Labs train on thousands of these attempts. Bought to improve models.

Only benchmarks use the Spatial Reasoning tag: the RL environment directories file their vendors under other categories. Browse all RL environments.

AI benchmarks 30

None confirmed yet.

30 not yet confirmedListed in the catalog; no line from its own site or paper confirms it yet.
  1. ArchSIBench
  2. MindEdit-Bench
  3. SpaceDG
  4. SPACENUM: Revisiting Spatial Numerical Understanding in VLMs
  5. CrossViewBench
  6. EscherVerse
  7. SceneFunRI
  8. SiT-Bench
  9. SpatialUAV
  10. CurveBench
  11. From Hallucination to Grounding
  12. Polaris-Bench
  13. QSTRBench
  14. Seeing Isn't Knowing
  15. SpatialBabel
  16. CogRail
  17. InterGPS
  18. PanoWorld
  19. SpatialAct
  20. STReasoner
  21. TangramPuzzle
  22. 3,034-sample human-curated spatial understanding benchmark
  23. Do LLMs Build World Models From Text? A Multilingual Diagnostic of Spatial Reasoning
  24. RefSpatialBench
  25. SUNRGBD
  26. Hypersim
  27. KnotBench
  28. NoReGeo: Non-Reasoning Geometry Benchmark
  29. RefCOCO-avg
  30. SpaRRTa

Next: pick another domain. Tags come from the RL Research daily scrape, canonicalized nightly.