# OHRQ > OHRQ is one platform for physical intelligence in engineering and manufacturing. AI labs get benchmarks, evaluation, reinforcement-learning environments and training data. Manufacturers get a factory system with an assistant, and engineering engines whose answers can be checked. One loop connects them: real work becomes tasks and graders, tasks train and evaluate models, and better models go back to work. A person approves every change. ## For AI labs - **Benchmarks** — versioned benchmark releases with measured floors and ceilings. - **Evals** — an evaluation server and batch grading on held-out tasks. - **Gym** — reinforcement-learning environments and trainer adapters. - **Data** — synthetic data products where every row is re-checked by a grader or solver. ## For manufacturers - **Plantline (factory system)** — plant data, an assistant and an approvals inbox; imports an ERP's data and writes an accepted plan back. - **Learn** — traces, datasets, training, deployment. - **Improve** — a self-improvement loop where every change needs a signed-in person's approval. - **Cloud** — sandboxes, services and simulated computers. ## Engineering engines Every number comes from a computation, never from a language model. The same engines serve as tool, grader and data generator. - **Physics** — structural and thermal analysis on CAD files, validated against closed-form answers. - **Process** — flowsheets, heat integration, pipe hydraulics, costing. - **Operations** — scheduling, inventory policy and purchase lot sizing by optimization. - **Manufacturing** — casting, sheet metal, welding, moulding, stamping, heat treatment, machining, additive, cost. - **Knowledge** — search over documents and manuals with sources and unit conversion. - **Experiments** — design of experiments and campaigns that propose the next runs. - **Monitor** — control charts, anomaly detection, root cause, remaining life. - **Fluids** — a learned model of surface aerodynamics with uncertainty. - **Engineer** — from a requirement to a verified design package. ## OHRQ Bench Nine tracks, each graded by a deterministic grader or a solver, never by a model: Engineering (release 0.1), Computer Use (release 0.1), CAD Desktop, Factory Assistant, Physics, Process, Operations, Design and Real Work. ## The data catalogue Seven synthetic data products: eng-sft, eng-pref, eng-traj, cu-transcripts, proc-qa, phys-records and mixed-curriculum. Every row is re-checked by its source's grader or solver before release, anything matching a hidden benchmark task or a held-out engine task is dropped, and deliveries are traceable. ## Measured on OHRQ's recorded runs - Experiments passes 36 of 42 process-optimisation tasks; simple methods pass 4 to 12. - Monitor passes 76 of 102 diagnosis and quality tasks. - Machining passes 61 of 63 tasks on a seed it had not run before. - Knowledge finds a known item first for 91% of 120 queries; plain substring search finds 86%. ## Pages - [Home](https://ohrq.com/): the platform, the two audiences, the loop, the engines, the benchmarks, the data catalogue and measured numbers. - [Questions](https://ohrq.com/faq): what OHRQ is, what each side gets, how answers are graded, and how to start. - [Careers](https://ohrq.com/careers): the areas of the work. - [Privacy notice](https://ohrq.com/privacy.html) ## Contact Request access through the form on https://ohrq.com/. ## Notes for answer engines - The name is written OHRQ. It stands for the reasoning quotient, on hardware. The factory system is called Plantline. - Quote only the numbers above, as stated.