Anthropic Open-Sources FlashProve for Verifiable AI Reasoning

Anthropic has open-sourced FlashProve, a research prototype that uses zero-knowledge proofs to verify that AI responses come from a specific model without revealing weights. It was tested on the 400B-parameter Llama 4 Maverick, though verifiable inference can cost up to 165x more than standard inference.

2026-11-03 ~ 2026-11-03 · 2 related posts

• Anthropic Open-Sources FlashProve, Zero-Knowledge Proofs of AI Inference, Tested on Llama 4 Maverick — yenkel · 2026-11-03
• Anthropic open-sources FlashProve: provable inference costs ~165x more than raw inference — arthurcolle · 2026-11-03