Latest / Frontier models and safety
Anthropic releases FlashProve system for cryptographic proof of AI inference
See how this story connectsAnthropic published FlashProve, an open-source prototype that uses zero-knowledge proofs (zkSNARKs) to prove that a committed set of model weights produced a given response. On a single H100 GPU, the system generates proofs at over 100 tokens per second for Llama 4 Maverick, a 400-billion-parameter mixture-of-experts model with 4-bit quantization, with proofs of about 19 MB per 1,024-token request. Proof generation takes about 70 to 95 times the GPU time of running the model on vLLM. Anthropic released executable code, Lean proofs and documentation under the Apache 2.0 licence.
Why it matters
Anthropic says such proofs could let users, regulators and companies check that a response came from a committed model without seeing its weights, which would help detect model substitution and contribute to verifying AI pacing requirements.
- When
- 3 Nov 2026
- ID
WTA-N-2026-0648- Where
- International
- Who
- Anthropic, EPFL, StarkWare, MIT, Succinct; Roy Rinberg, Alex Allain, Alessandro Chiesa, Miranda Christ, Rahul Dalal, Giacomo Fenzi, Ben Mayer, Guy Weissenberg
Summary written with AI by wheresthe.ai, from Anthropic and the reporting credited above. Checked against the source on 3 Nov 2026.
Source copy: saved on 3 Nov 2026 to check this story and not republished, because the text belongs to its publisher. Also held by the Internet Archive.
Cite this story
WTA-N-2026-0648: “Anthropic releases FlashProve system for cryptographic proof of AI inference.” wheresthe.ai, 3 Nov 2026. https://wheresthe.ai/d/intl-anthropic-releases-flashprove-system-for-cryptographic-2026-11/ (source: Anthropic).
Spotted an error? Email [email protected].