Coding

Environments that train and evaluate models on real software engineering: resolving issues in production repositories, writing and reviewing pull requests, debugging failing test suites, and shipping features end-to-end. Coding is the most mature RL environment category: benchmarks like SWE-bench proved that verifiable rewards from test suites scale, and frontier labs now buy coding environments by the thousand.

29 startups

NameDomainsLocationTeamFundingRaising
AfterQuery
Expert human data and RL environments across code, finance, and computer use
Multi-Domain, Coding, FinanceSan Francisco, New York, Seattle51-100$30.5M total (reported)-
Akhara
Enterprise and code RL environments
Enterprise, CodingSan Francisco1-10--
BenchFlow
Open-source benchmark hub and eval infrastructure for agents
Enterprise, Browser, CodingSan Francisco1-10~$1M (reported)-
Bespoke Labs
Data curation and RL environment recipes from ex-Google DeepMind researchers
Coding, Machine LearningMountain View, Menlo Park, Bangalore, San Francisco11-25~$40M (reported)-
Cua
Open-source computer-use agent infrastructure and environments
Coding, Computer Use, Agents InfrastructureSan Francisco1-10Pre-seed/Seed (YC X25)-
Datacurve
Frontier coding data and repository RL environments via the Shipd bounty platform
Coding, RLHFSan Francisco26-50Series A, $15M led by Chemistry (Oct 2025); $17.7M total-
Deeptune
Code and computer-use environments; acquired by Mercor
Coding, Computer UseNew York26-50Series A, $43M; acquired by Mercor (2026)No
E2B
Open-source cloud sandboxes for AI agents
Agents Infrastructure, CodingSan Francisco26-50Series A, $21M (Insight Partners, July 2025); $32M total-
Emulated
Code and ML RL environments
Coding, Machine LearningSan Francisco1-10--
Exabite
Code RL environments with realistic software execution
CodingRemote1-10--
Habitat
Code and desktop interaction RL environments
Coding, Computer UseNew York1-10--
HUD
Evals and RL environments platform for computer-use agents
Computer Use, Coding, Long Horizon, Agents Infrastructure, Custom EnvironmentsSan Francisco, Singapore11-25$15M raised (YC W25, Exceptional Capital)-
Huzzle Labs
Long-horizon code, tool-use, and enterprise workflow environments
Long Horizon, Coding, EnterpriseLondon, Berlin, San Francisco26-50~$6M (reported)-
Idler
Code environments with realistic execution constraints
CodingSan Francisco11-25--
Mechanize
RL environments to automate software engineering, founded by ex-Epoch AI researchers
CodingSan Francisco51-100~$9.1M (reported)-
Metaphi
Code and enterprise RL environments
Coding, EnterpriseSan Francisco, New York1-10--
Originator
Long-horizon computer-use environments
Computer Use, Coding, Long HorizonLondon, Paris1-10--
pre.dev
Software-planning platform offering coding and long-horizon RL environments
Coding, Long HorizonDelaware1-10--
Preference Model
Stealth startup working on preference and reward modeling
Machine Learning, CodingSan Francisco, Toronto, Seattle11-25--
Proximal
Long-horizon coding RL environments built from real codebases
Coding, Long HorizonSan Francisco, Bangalore26-50--
Quesma
Security-domain RL environments and binary analysis evals
Cybersecurity, CodingWarsaw11-25--
ReasonCore
Science and code reasoning environments and benchmarks
Science, CodingSan Francisco11-25--
Refresh
Simulation engines with verifiable rewards for coding and computer use
Coding, Computer Use, SimulationSan Francisco1-10--
Rise Data Labs
US-based expert human data and custom RL environments for enterprise AI
Computer Use, Coding, Finance, Cybersecurity, Legal, Data Labeling, Enterprise, Multi-Domain, RLHF, Custom EnvironmentsUnited States--Yes
Runloop
Devboxes for training and evaluating coding agents
Coding, Agents InfrastructureSan Francisco1-10Seed, ~$7M-
Scale
Data-labeling incumbent extending into agent evals and RL environments
Multi-Domain, Coding, Data Labeling, RLHFSan Francisco, New York, Washington DC, London250+$1.6B+ raised; Meta invested $14.3B at ~$29B valuation (June 2025)-
Snorkel
Programmatic data platform expanding into expert evals and RL environments
Multi-Domain, Coding, Machine Learning, Data Labeling, RLHFSan Francisco, Redwood City, New York101-250Series D, $100M at $1.3B valuation (2025)-
Turing
AGI infrastructure: coding data and RL environments at scale
Coding, Multi-Domain, Data Labeling, RLHFSan Francisco, Palo Alto, Gurugram250+Series E, $111M at $2.2B valuation (2025)-
Vetto AI
Code and computer-use environments from ex-DeepMind/Instagram founders
Coding, Computer UseSan Francisco, São Paulo, London1-10--

Other domains