{"rewrite":{"id":"r_0c41d5b67174efc018cba7ec","clusterId":"c_5328932ac91736c252832083","slug":"anthropic-builds-flashprove-to-prove-which-model-produced-an-output","model":"deepseek-v4-1-flash","headline":"Anthropic Builds FlashProve to Prove Which Model Produced an Output","summary":"Anthropic has developed FlashProve, a system that lets a service provider prove an output came from a specific AI model. It uses zkSNARK proofs that can be generated more cheaply than with existing methods and that omit model weight information, so closed models can be covered without exposing them. Verification takes a few seconds on a 24-core CPU. Anthropic published the code as an unmaintained research prototype on GitHub, with a technical report.","whyItMatters":"Cheap, weight-hiding proof that a given model produced a given output would close the gap that lets a provider advertise one model and serve another.","webCardHtml":"\u003cp\u003eThe stated problem is a mismatch between the model a service advertises and the model it runs. Anthropic lists three ways that gap gets exploited: a provider quietly serves a cheaper model for the same price, a provider swaps in a different model after an auditor has cleared the original, or an attacker impersonates a model to reach users. FlashProve is Anthropic\u0026#39;s answer, built on zkSNARK proofs. The proof file carries no information about the model\u0026#39;s weights, so a provider can hand over proof for a closed model without revealing what is inside it. Verification finishes in seconds on a 24-core CPU. The published repository is labeled a research prototype and is not maintained.\u003c/p\u003e","blueskyPost":"FlashProve's zkSNARK proofs omit model weight data, which is why closed models can be verified without being opened up. That design choice is what makes the system usable outside Anthropic.","twitterPost":"FlashProve works on closed models because its proofs omit model weight data. Verification runs in seconds on a 24-core CPU.","threadsPost":"FlashProve's design keeps model weight data out of the proof entirely, so a provider can show which model made an output without opening the model itself. That is what lets the system reach closed models at all.","newsletterBlurb":"Anthropic has developed FlashProve, a zkSNARK-based system that proves an output was generated by a specific AI model without revealing model weights. Proof verification takes a few seconds on a 24-core CPU. The code is published on GitHub as a research prototype, with a technical report alongside it.","attributionJson":"[{\"source\":\"GIGAZINE\",\"url\":\"https://gigazine.net/news/20261104-anthropic-flashprove/\",\"title\":\"Anthropic Develops \\\"FlashProve,\\\" a System That Can Prove \\\"Output Was Generated by a Specific AI Model\\\"\"}]","lintFlagsJson":null,"lintHits":0,"costUsd":0,"inputTokens":5510,"outputTokens":694,"status":"published","repairAttempts":0,"nextRepairAt":null,"factsAttemptedAt":1793760607,"createdAt":"2026-11-04T02:46:10.000Z","publishedAt":"2026-11-04T02:48:56.000Z","updatedAt":"2026-11-04T02:48:56.000Z"},"cluster":{"id":"c_5328932ac91736c252832083","canonicalTitle":"「特定のAIモデルで出力が生成されたこと」を証明できる仕組み「FlashProve」をAnthropicが開発","representativeArticleId":"a_620db0b6d48e84a8718125d7","sourceCount":1,"writtenSourceCount":1,"writeAttempts":0,"isSolo":true,"entitiesJson":"{\"anime_titles\":[],\"manga_titles\":[],\"work_titles\":[\"FlashProve\"],\"studios\":[\"Anthropic\"],\"people\":[],\"type\":\"news\",\"domain\":\"other\",\"is_roundup\":false}","contentType":"news","status":"published","firstSeenAt":"2026-11-04T02:07:00.000Z","lastSeenAt":"2026-11-04T02:07:00.000Z","updatedAt":"2026-11-04T02:48:52.000Z"},"attribution":[{"source":"GIGAZINE","url":"https://gigazine.net/news/20261104-anthropic-flashprove/","title":"「特定のAIモデルで出力が生成されたこと」を証明できる仕組み「FlashProve」をAnthropicが開発"}],"entities":{"anime_titles":[],"manga_titles":[],"work_titles":["FlashProve"],"studios":["Anthropic"],"people":[],"type":"news","domain":"other","is_roundup":false},"keyFacts":["Anthropic developed FlashProve to prove that an output was generated by a specific AI model.","FlashProve uses the zkSNARK cryptographic proof method, which does not include information about model weights.","Verification of a FlashProve proof file completes in a few seconds on a 24-core CPU.","Anthropic published the FlashProve code on GitHub as a research prototype that is not maintained.","A FlashProve technical report is published as a PDF."],"pageLinks":{"healthSlugs":["flashprove","anthropic"],"storySlugs":[],"entities":[{"slug":"flashprove","supported":0,"disputed":0,"contradicted":0,"unverified":2,"overall":"unverified","story_count":1,"has_dossier":false},{"slug":"anthropic","supported":18,"disputed":0,"contradicted":0,"unverified":39,"overall":"green","story_count":41,"has_dossier":true}],"stories":[],"entitiesOk":true,"storiesOk":true}}