---
title: "Anthropic releases FlashProve to cryptographically verify which model produced a reply"
url: https://www.parallelquant.com/posts/anthropic-releases-flashprove-to-cryptographically-verify-which-model-pr-4fb264
source_name: "MarkTechPost"
source_url: https://www.marktechpost.com/2026/11/03/anthropic-flashprove-provable-inference-zksnark/
published: 2026-11-03T22:08:22.000Z
topics: ["research", "llms"]
publisher: "Parallel Quant"
---

# Anthropic releases FlashProve to cryptographically verify which model produced a reply

*2026-11-03 · Source: [MarkTechPost](https://www.marktechpost.com/2026/11/03/anthropic-flashprove-provable-inference-zksnark/)*

Anthropic released FlashProve, an Apache 2.0 zkSNARK (zero-knowledge proof) prototype that proves a reply came from a committed model without revealing the weights. It proves Llama 4 Maverick at about 110 tokens per second on one H100 GPU, with proofs around 19 MB.

**Why it matters:** Users of API resellers and third-party hosts currently have no way to know whether they get the model they pay for, so verifiable inference could curb silent model substitution. It also fits a broader push on provenance and trust, alongside the shared scraping opt-out standard and recent supply-chain incidents. Proof size and overhead still limit practical use for now.

**Topics:** research, llms

---
Read the original: https://www.marktechpost.com/2026/11/03/anthropic-flashprove-provable-inference-zksnark/
Canonical: https://www.parallelquant.com/posts/anthropic-releases-flashprove-to-cryptographically-verify-which-model-pr-4fb264
Published by Parallel Quant — https://www.parallelquant.com