Verde is a protocol from the company Gensyn for checking machine-learning jobs, such as inference, fine-tuning or training, that a client hands to untrusted compute providers.
Several providers run the same job. If their results differ, halving the disputed computation repeatedly (a bisection game) narrows the dispute to one operation, which a referee re-runs to decide who is right. The client gets the correct result if at least one provider is honest.
This needs identical results on different hardware, which Gensyn reports its RepOps library provides by fixing the order of floating-point operations. Gensyn reports settling markets in its Delphi app with REE, a runtime built on RepOps whose receipts let anyone re-run the inference, and states that Verde runs in its Judge service. RepOps ships as closed binaries.
In Gensyn's paper, RepOps roughly doubled each provider's Llama-8B inference time. As of September 2026 no independent evaluation has been published.
Gensyn reports using its public runtime in production to settle markets whose final judge computation anyone can re-run, but all evidence of that use comes from Gensyn, and no independent security evaluation of Verde or RepOps has been published.
Assessed use: reproducing declared-model inference from receipts in Gensyn's information-market service
On this page
What it is
Verde is a protocol from Gensyn for checking machine-learning programs, such as inference, fine-tuning and training of language models, that a client delegates to untrusted compute providers 1. It adapts refereed delegation, a cryptographic technique in which a client sends the same job to several providers and a referee settles disagreements 1. RepOps (Reproducible Operators) is the library the Verde paper introduces to make honest providers' results bitwise identical on different hardware 1. Most of the paper's authors work at Gensyn 1.
Verde relies on bit-exact inference, and for training jobs it resolves disputes over checkpoints, as in training-transcript verification. Gensyn reports that Verde and RepOps are deployed in Judge, its service for verifiable AI evaluation 2 3. It also reports settling markets in Delphi, its information-market app, with REE 5.
How it works
- The client gives the same job to two or more providers 1, and arbitration starts only if their outputs differ 2.
- A two-level bisection game finds the first step, and then the first operation in the computational graph, on which the providers disagree 1 2.
- The referee re-runs that one operation to decide which output is correct. The authors state that this cuts the referee's communication and compute by two orders of magnitude compared with re-running the whole disputed step 1.
RepOps fixes the order of floating-point operations for common operators such as matrix multiplication, so that an honest provider's result does not depend on its GPU 1 2. Gensyn's Reproducible Execution Environment (REE) packages this for large language model (LLM) inference. It writes a receipt holding the model, prompt, output and the metadata needed to reproduce the run 4. Its reproducible mode uses RepOps kernels. A separate deterministic mode uses PyTorch's deterministic algorithms and reproduces results only on the same hardware 6.