Understand this story
On November 2, 2026, Anthropic released and open-sourced FlashProve, a cryptographic verification system that lets developers determine whether a given output was generated by a committed model without needing the model weights. According to Sina Tech, the system runs cp_verify on an ordinary CPU and completes verification in about 17 seconds; before launch, it takes about 39 minutes on a single H100 to generate a commitment ID for the weights, each request generates a proof of about 19 MB, and clients can verify offline. The test subject was the 400-billion-parameter Llama 4 Maverick; security relies on the BLAKE2 hash function, and 95.2% of the project code consists of Lean 4 formal proofs. Another report noted that the GPU time required to generate proofs is about 70 to 95 times that of ordinary vLLM inference; currently it supports only 1024-token context, integer models, and greedy decoding, and it proves the provenance of an answer rather than the correctness of its content.
AI-generated from public reports
Latest development · Report
Anthropic Open-Sources FlashProve, Using Zero-Knowledge Proofs to Verify Model IdentityDevelopments
1 confirmed developments
Anthropic releases open-source research prototype FlashProve
Source:Sina Tech
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Nov 4
36Kr · AI
Anthropic Open-Sources FlashProve, Using Zero-Knowledge Proofs to Verify Model Identity
Anthropic has open-sourced the FlashProve system, which is based on zero-knowledge proofs and allows model service providers to prove that a given answer was generated by a specified model without disclosing the weights. Tests were conducted on the 400B-parameter Llama 4 Maverick: a single NVIDIA H100 can generate proofs for about 110 tokens per second, the proof file for each 1024-token request is about 19 MB, and verifying one proof on a 24-thread CPU takes about 17 seconds. The GPU time to generate proofs is about 70 to 95 times that of ordinary vLLM inference. Currently it only supports a 1024-token context, integer models, and greedy decoding, and it only proves the source of an answer rather than the correctness of its content.
Ayi AI Notes
Anthropic Open-Sources FlashProve: Verifying Model Outputs with Zero-Knowledge Proofs
Anthropic has announced its verifiable inference research FlashProve, enabling developers to verify whether a given response came from a specified model, and has open-sourced the code. The solution is based on a zkSNARK architecture: model outputs come with a mathematical proof attached, and users can complete verification on an ordinary CPU within seconds without leaking model parameters. According to reports, on a 400B MoE Llama 4, a single H100 can generate proofs for 110 tokens per second, with proof files of about 19MB; the code and the Lean formal security proof are open-sourced under the Apache 2.0 license.
Sina Tech
Anthropic Open-Sources FlashProve: Verifying the Authenticity of Model Outputs in 17 Seconds
Anthropic has announced and open-sourced FlashProve, a cryptographic verification system that lets developers determine whether a given output was generated by a committed model in about 17 seconds by running cp_verify on an ordinary CPU, without needing the model weights. The system has three steps: before deployment, a commitment ID for the weights is generated on a single H100 in about 39 minutes; each request generates a proof of about 19MB, which the client verifies offline. The test subject is the 400-billion-parameter Llama 4 Maverick; security relies on the hash function BLAKE2, and 95.2% of the project's code consists of Lean 4 formal proofs.
Sina Technology
Anthropic Open-Sources FlashProve, Using Zero-Knowledge Proofs to Verify Model Identity
Anthropic has open-sourced the FlashProve system, which uses zero-knowledge proofs to let model providers prove that a given response was generated by a specified model without disclosing the weights. The system was tested on Meta Llama 4 Maverick, a 400B-parameter MoE model: a single NVIDIA H100 can generate inference proofs for about 110 tokens per second, the proof file for each 1024-token request is about 19 MB, and CPU verification of one proof takes about 17 seconds. Anthropic also notes that the GPU time required to generate proofs is about 70 to 95 times that of ordinary vLLM inference; currently only 1024-token contexts, integer models, and Greedy Decoding are supported, and what is proven is the source of the response rather than the correctness of its content.
TechCrunch AI
Anthropic develops FlashProve system to cryptographically prove AI model identity
Anthropic is developing an open-source system called FlashProve that uses cryptographic proofs to verify that a given output actually came from a specific model. The system requires AI labs to publish a fingerprint for each model that does not expose its weights, which can prevent vendors from secretly swapping in a cheaper or more expensive model, and can also verify third-party claims that a task was completed by a certain model; by design it is lightweight enough to allow verification to be completed on local devices. The researchers have published the technical specification and part of the open-source code, and the research is still ongoing, with no plans for deployment at Anthropic or other labs at this time.
MarkTechPostRelated discussion
MarkTechPost joined this discussion
This post contributes to heat; it is not a formal report or confirmed fact.
Rohan Paul
Anthropic open-sources FlashProve, which can verify on an ordinary CPU which model a response came from
Anthropic has open-sourced FlashProve, enabling servers to mathematically prove that a given response came from a specified model, with verification completable on an ordinary CPU. On Meta's Llama 4 Maverick (400B parameters), a single Nvidia H100 generates proofs at 110 tokens/second, each about 19MB, and an ordinary CPU verifies one in about 17 seconds; the only previously known system capable of proving this model, VerInf, generated 93.6GB proofs, with verification taking nearly a day. The system was built by Claude, and because the code is difficult to audit, its security guarantees are mathematically proven in Lean 4.
Anthropic
Anthropic releases FlashProve, which can verify the source of model responses and open-sources the code
Anthropic has announced FlashProve, a new research direction in verifiable inference. Built on existing cryptographic techniques, it lets developers verify whether a given response came from a specified AI model, and Anthropic says it is significantly more efficient than previous open-source methods. The proof can be used to make it harder for attackers to hijack conversations, and it also allows safety evaluators to confirm that the model being tested is the same as the model serving the public. Anthropic has open-sourced the FlashProve code and lists collaborators including Alessandro Chiesa.
Nov 3
Unite.AI
Anthropic Releases FlashProve, Using Cryptography to Prove the Provenance of Model Reasoning
On November 2, 2026, Anthropic released FlashProve, an open-source research prototype that uses zkSNARKs to prove that a given response came from committed model weights without revealing the weights themselves; for Meta's Llama 4 Maverick, it reports a proof generation speed of 110 tokens/s. A single H100 plus a 24-core CPU can prove 24 requests in about 223 seconds, generating an 18.78 MB proof per request, with the offline commitment phase taking about 39 minutes; after downloading a commitment of about 6.5 MB, the client verifies on CPU in 17 seconds with 24 threads and 21 seconds single-threaded. Compared with FP8 inference running vLLM on 8 H100s, the single-GPU time required for proving is about 72 times higher, and about 94 times higher under maximum load. The code is released under Apache 2.0, and version 0.16.0 is labeled as an unmaintained research artifact with no independent security audit yet; proofs cover only the integer-weight version under greedy decoding, with a context limit of 1024 tokens.
Heat trend for this story
- Current
- 61
- Period peak
- 71 Nov 4 09:00
- 24-hour change
- —
Move the pointer or select the chart to inspect hourly heat. Use the left and right arrow keys with a keyboard.