image-understanding
University of California, Berkeley
3D-simulated household tasks driven by language instructions and egocentric video - the visual sibling of ALFWorld.
Google DeepMind
57 Atari 2600 games played from raw pixels - the foundational reinforcement-learning benchmark from DeepMind's DQN era.
6,141 multimodal math problems combining diagrams, charts, and figures - the canonical "math + vision" benchmark.
Carnegie Mellon University
910 visually grounded web tasks across three self-hosted sites (Classifieds, Shopping, Reddit) requiring image understanding to complete.
MIT CSAIL
Aligned text-and-3D embodied environment - agents learn household tasks (pick & place, heat, cool, clean) as both TextWorld games and visually-rendered ALFRED scenes.
Meta Platforms
OpenEnv wrapper around the Arcade Learning Environment - the classic 57-game Atari 2600 benchmark suite that defined deep RL.
ServiceNow Research
ServiceNow's unified Gym-style framework for web agents - wraps WebArena, MiniWoB, VisualWebArena, WorkArena, AssistantBench, WebLINX, and more under one Playwright-backed interface.
MathVista multimodal mathematical reasoning environment with vision and structured output requirements
Numina
An 860k-problem competition-math dataset with detailed solutions, the open community's go-to corpus for training math-specialized LLMs.