Anthropic Releases FlashProve Open-Source Prototype Using Zero-Knowledge Proofs to Verify LLM Outputs

According to PANews, on November 3, Anthropic released FlashProve, an open-source research prototype that uses zero-knowledge proofs to verify that model responses were generated by specified model weights without exposing the weights themselves. During testing on the 400-billion-parameter Llama 4 Maverick model, a single H100 GPU generated proof at approximately 110 tokens per second, though GPU time required was 70 to 95 times longer than standard inference. The prototype currently supports up to 1,024 tokens of context and remains in research stage.
Disclaimer: The information on this page may come from third-party sources and is for reference only. It does not represent the views or opinions of Gate and does not constitute any financial, investment, or legal advice. Virtual asset trading involves high risk. Please do not rely solely on the information on this page when making decisions. For details, see the Disclaimer.
Comment
0/400
No comments