RL Environment and Expert Data Market Map, November 2026
Patronus AI TeamAs of November 20266 min read
A dated snapshot. Patronus AI drew this map and is one of the companies on it. Groups and companies are alphabetical, not ranked or sized. Each placement follows the company’s own description, linked below. Checked Nov 11, 2026.
The RL environment and expert data market splits into seven groups as of November 2026. Expert data and evaluation companies, research labs that build environments, builders of software replicas and simulated workplaces, domain specialists, game environments, open hubs and platforms, and the agent infrastructure underneath. This map places 23 companies in those groups.
The companies are the ones in our RL environment companies dataset, which gives each one’s founding year, headquarters and public work with sources. Download it there as CSV or JSON.
Snapshot as of November 2026. Last updated
The map
Each box is one kind of company. Within a box, companies are alphabetical.
The companies on the map
The same companies, with what each says it builds and a link to the page that says it.
Agent infrastructure
Sandboxes and browsers that agents run in. They supply the machine, not the tasks or graders.
- Browserbase Browserbase runs cloud browsers for AI agents, and its BrowserEnv wraps those sessions as an environment to train and evaluate browser agents. src
- E2B E2B runs isolated microVM sandboxes for AI agents, in its cloud or the customer's, for coding, computer use, background jobs and RL rollouts. src
Domain specialists
Environment builders that go deep in one field.
- Halluminate A data research lab building benchmarks and RL environments for knowledge work, starting with financial services and consulting. src
- Mechanize Mechanize builds RL environments and evals in which frontier coding agents do software engineering tasks scored by a grader. src
Expert data and evaluation companies
Companies built on human expert data and evaluation that also sell RL environments.
- Invisible Technologies Builds expert-built RL environments where domain experts define task logic, design reward rubrics and annotate trajectories, alongside expert data. src
- Labelbox Labelbox builds RL training environments and evals for post-training on reasoning, tool use and computer use, co-designed with domain experts. src
- Mercor Mercor supplies expert human work for frontier research, AI benchmarks and AI agent training for AI labs and enterprises. src
- Scale AI Scale AI's Enterprise Atlas RL environments train agents on long-horizon work across real enterprise systems and desktop environments. src
- Snorkel AI Snorkel AI develops specialized training data and environments for frontier labs and AI teams' models and agents. src
- Surge AI Surge AI sells off-the-shelf training data, evals and RL environments built by the experts behind its frontier-lab work. src
- Turing Turing builds iterable UI and non-UI RL environments, including app replicas and MCP backends, for agents to learn long-horizon workflows. src
Game environments
Environments built from games with clear rules and scores.
- Good Start Labs Builds reinforcement learning environments and world-model data from games, plus public benchmarks that measure model capability in games. src
Open hubs and platforms
Tools and shared environments for teams that build their own.
- HUD A platform and SDK for building RL environments, turning expertise into environments that train and evaluate models, with QA agents auditing traces. src
- Prime Intellect Prime Intellect offers an open-source library and hub for building and sharing RL environments, plus hosted RL training, evals and sandboxes. src
Research labs that build environments
Companies that publish research and benchmarks and build environments from it.
- Andon Labs Andon Labs studies frontier AI safety by running AI-run businesses in the real world and comparing models in simulated long-horizon evals. src
- Bespoke Labs Bespoke Labs builds RL environments and infrastructure so frontier labs and enterprises can train long-horizon agents for production. src
- General Reasoning An AI research company training models for long-horizon work, with environment design as a research area and OpenReward, an open platform serving RL environments. src
- Patronus AI Patronus AI is a behavioral research lab for AI that studies where models fail and builds simulated environments and training data to improve them. src
Software replicas and simulated workplaces
Builders of cloned apps, websites and whole simulated workplaces for agents to practice in.
- Chakra Labs Chakra Labs is a research and design lab building expert-authored tasks and pixel-perfect environments to train models and evaluate real-world capability. src
- Fleet Fleet builds training gyms for agents, high-fidelity simulation environments where AI practices tasks and humans provide supervision. src
- Idler A frontier data lab building RL environments from real production work, across frontier coding, economically valuable workflows and bespoke evaluations. src src
- Plato Plato builds simulated worlds of the software and tasks of real jobs, where models train and are evaluated on that work. src
- Veris AI Veris AI builds a simulated copy of a company's systems, data, APIs and users where AI agents are tested and trained. src
How the map was made
We placed each company by what it says it builds on its own site, read on Nov 11, 2026, and gave each company one group, the one its main product fits. Many companies do more than one thing, and the line under each name says what. Groups follow our RL environment startups guide, with expert data companies and agent infrastructure added.
The map does not size, rank or score anyone, and it leaves out funding and customers, which change often. To compare the companies, see our guide to RL environment companies. To judge an environment, see our guide on how to evaluate RL environments.
Frequently asked questions
What does the RL environment market look like in 2026?
As of November 2026 it has seven kinds of companies. Expert data and evaluation companies, research labs that build environments, builders of software replicas and simulated workplaces, domain specialists, game environments, open hubs and platforms, and agent infrastructure such as sandboxes and browsers.
Which companies build RL environments?
This map places 23, listed alphabetically within each group. They include Andon Labs, Bespoke Labs, Chakra Labs, Fleet, General Reasoning, Good Start Labs, Halluminate, HUD, Idler, Mechanize, Patronus AI, Plato, Prime Intellect and Veris AI, and expert data companies such as Invisible Technologies, Labelbox, Mercor, Scale AI, Snorkel AI, Surge AI and Turing.
How are expert data companies different from RL environment builders?
Expert data companies are built around human expert data and evaluation and sell environments next to it. Environment builders and research labs sell environments as their main product, and hubs and infrastructure companies sell the tools to build and run them.
Where does the data behind the map come from?
From our RL environment companies dataset, where every value links to the company’s own page it was read from. The dataset downloads as CSV and JSON.
Patronus AI · November 2026 snapshot · More resources