Prime Intellect

An AI compute, training and inference company; developer of TOPLOC, a hashing scheme for verifying LLM inference.

www.primeintellect.ai

Prime Intellect describes itself as offering an integrated compute, training, inference and sandbox stack for AI models 1. Its verification work centres on TOPLOC, a form of sampled inference recomputation:

  • Seven of the eight authors of the ICML 2025 paper that introduces TOPLOC are at Prime Intellect 2. An inference provider using it records the largest activations of the last hidden layer, so that a verifier who re-runs the sequence can check the claimed model, prompt and precision 2.
  • Prime Intellect publishes TOPLOC as an MIT-licensed Python package 3, and reports integrations with vLLM and a fork of SGLang 4.
  • Prime Intellect reports that each inference worker in its INTELLECT-2 decentralized training run generated a TOPLOC proof for every sequence 5. It reports that a later version, TOPLOC v2, extends the scheme to pipeline-parallel inference and adds checks of token sampling 6. It reports using TOPLOC v2 to check the inference workers in its SYNTHETIC-2 data-generation run, on more than 1,250 GPUs, with a false-positive rate of 0.000925% over 4 million samples 7.
On this page

Implementations

Implementations this organization develops.

  • TOPLOC is a hashing scheme from Prime Intellect that lets a verifier check whether an inference provider ran the model, prompt and precision it claims.

Mechanisms and implementations whose records cite or describe this organization's work.

  • A verifier re-runs a random sample of an AI provider's logged queries on a trusted copy of the declared model and checks the outputs match.

Publications

Sources this organization authored or published.

Sources

  1. BPrime Intellect homepage. Prime Intellect. Source recordSupports: integrated compute, training, inference and sandbox stack
  2. AJ. M. Ong et al. (2025). TOPLOC: A Locality Sensitive Hashing Scheme for Trustless Verifiable Inference. Proceedings of the 42nd International Conference on Machine Learning (PMLR 267), pp. 47196-47211. Source recordSupports: TOPLOC paper; seven of the eight authors at Prime Intellect; what the method checks
  3. BPrime Intellect (2025). PrimeIntellect-ai/toploc (GitHub repository). GitHub. Source recordSupports: open-source TOPLOC package (MIT licence)
  4. CPrime Intellect (2025). TOPLOC: A Locality Sensitive Hashing Scheme for Trustless Verifiable Inference (blog post). Prime Intellect blog. Source recordSupports: vLLM and SGLang integrations (provider-reported)
  5. BPrime Intellect Team et al. (2025). INTELLECT-2: A Reasoning Model Trained Through Globally Decentralized Reinforcement Learning. arXiv. Source recordSupports: TOPLOC proofs in the INTELLECT-2 training run (provider-reported)
  6. CPrime Intellect (2025). SYNTHETIC-2. Prime Intellect blog. Source recordSupports: TOPLOC v2 extends the scheme to pipeline-parallel inference and sampling checks (provider-reported)
  7. CPrime Intellect (2025). SYNTHETIC-2 Release: Four Million Collaboratively Generated Reasoning Traces. Prime Intellect blog. Source recordSupports: TOPLOC v2 checks in the SYNTHETIC-2 run: 1,250+ GPUs, false-positive rate over 4 million samples (provider-reported)

Search

Full search page