General Reasoning

General Reasoning offers AI labs and developers API-hosted reinforcement-learning environments, plus infrastructure to deploy environments and train or evaluate agents at scale.

Sources:gr.inc/openreward gr.inc/careers/mts-infrastructure

gr.inc · updated Sep 29, 2026

RL environments Evals and benchmarks RL infrastructure

Add to compareSuggest an edit

Headquarters: London, United Kingdom
Source:gr.inc
Founded: 2025
Source:gr.inc
Headcount: 11-50
Source:linkedin.com
Total funding: $10.9M
Sources:pitchbook.com formds.com
Latest round: Seed, $10.9M, July 2025
Sources:pitchbook.com sec.gov

Trust & data

How the vendor protects buyer material, where its data comes from, and where it can run.

Deployment: Vendor-hostedCustomer cloud / VPC
Sources:gr.inc/openreward gr.inc/releases/introducing-openreward
Data sources: Synthetic / generated
Source:gr.inc
Quote
Synthetic / generated: Agents are placed in a sequential simulation of the 2023–24 English Premier League season
Stated confidentiality practices: It says agent sandboxes use isolated containers; its infrastructure materials also specify hardened container isolation for untrusted code, network policies, and tool-invocation audit logging.
Sources:gr.inc/careers/mts-infrastructure gr.inc/openreward
Not publicly listed: SOC 2, Trust center, Other certifications, Exclusivity

What they sell

Offerings: RL environments Evals and benchmarks RL infrastructure
Source:gr.inc
Domains: Multi-domain
Source:gr.inc
Delivery formats: Docker containersAPIManaged service
Sources:gr.inc/careers/mts-infrastructure gr.inc/openreward
Public samples & benchmarks: KellyBench
NFLBench
PaperBench
DCS
PRBench
Sources:gr.inc openreward.ai/GeneralReasoning/NFLBench openreward.ai/GeneralReasoning/PaperBench openreward.ai/GeneralReasoning/DCS
Named customers: MetaGoogle DeepMindQwenNebiusSarvam
Source:gr.inc

Every change to these facts is kept in the public claims log. Record co_67c095c3bb73.