Scale AI offers Scale RL Environments
Product
Simulated applications with expert-curated artifacts and verifiers for training AI agents on long-horizon professional workflows.
- Currently Offered
Simulated applications with expert-curated artifacts and verifiers for training AI agents on long-horizon professional workflows.
Scale RL Environments provides simulated computer and API-based applications for AI model developers training and evaluating agents. The product combines expert-curated professional artifacts with tasks, rubrics, and automated verifiers.
Its environment types include web applications, desktop virtual machines, and MCP servers. Teams use these environments to train computer-use and tool-use agents, inspect trajectories and rewards, and reset application state between runs.
Cua Environments Catalog offers scoped trajectory data packages for computer-use agents.
Redpine RL Environments train and evaluate agents on expert tasks from licensed data.
Cua Environments Catalog offers scoped trajectory data packages for computer-use agents.
Redpine RL Environments train and evaluate agents on expert tasks from licensed data.
Digital World Models predict and simulate agent actions in digital workflows for AI training.
Expert-curated datasets and reinforcement learning environments for training and evaluating AI models.
Isolated environments for reinforcement learning, agent tool use, and model evaluation on GPU infrastructure.
An integrated compute, training, inference, and sandbox stack for enterprises to train and deploy their own AI models.
An open-source asynchronous reinforcement learning framework for training large language models at scale.
Open-source Python framework for building sandboxed reinforcement learning environments that run and score AI agents.
Enterprise product that runs AI simulations of customer and population behavior.
A benchmark that evaluates AI agents on end-to-end, long-horizon software engineering workflows in production-grade systems.
Agent Arena is a public leaderboard ranking developer tools by autonomous AI agent adoption.
Reinforcement learning environments and evaluations for AI chip design tasks.
Observability and evaluation platform for shipping quality AI agents at scale.
Train models on B200s, T4s and inference-optimized infra with one-click deploy to production.
Tools for building, deploying, evaluating, and governing enterprise AI and machine learning systems.
Foundation models and AI agents for spatial-temporal reasoning in virtual and physical environments.
Digital World Models predict and simulate agent actions in digital workflows for AI training.
Expert-curated datasets and reinforcement learning environments for training and evaluating AI models.
Isolated environments for reinforcement learning, agent tool use, and model evaluation on GPU infrastructure.
An integrated compute, training, inference, and sandbox stack for enterprises to train and deploy their own AI models.
Applied Compute's Agent Cloud trains enterprise-specific AI agents.
An open-source asynchronous reinforcement learning framework for training large language models at scale.
Open-source Python framework for building sandboxed reinforcement learning environments that run and score AI agents.
Enterprise product that runs AI simulations of customer and population behavior.
A benchmark that evaluates AI agents on end-to-end, long-horizon software engineering workflows in production-grade systems.
Agent Arena is a public leaderboard ranking developer tools by autonomous AI agent adoption.
River API lets developers fine-tune and run reinforcement learning on open-weight models.
Reinforcement learning environments and evaluations for AI chip design tasks.
Fugu is a multi-agent orchestration API that routes complex tasks across a pool of frontier language models.
Realm is micro1's data engine of RL environments and expert production.
Observability and evaluation platform for shipping quality AI agents at scale.
Train models on B200s, T4s and inference-optimized infra with one-click deploy to production.
Tools for building, deploying, evaluating, and governing enterprise AI and machine learning systems.
Foundation models and AI agents for spatial-temporal reasoning in virtual and physical environments.
