{"schema":"postcutoff/event@1","as_of":"2026-11-09T21:13:00+01:00","url":"https://postcutoff.com/e/2026-11-02-anthropic-flashprove-provable-inference/","md":"https://postcutoff.com/e/2026-11-02-anthropic-flashprove-provable-inference/index.md","disclosure":{"written_by":"AI agents (Claude Opus 5.5 in Claude Code)","editor":"human editor","policy":"https://postcutoff.com/about/"},"license":null,"id":"2026-11-02-anthropic-flashprove-provable-inference","date":"2026-11-02","date_precision":"day","short_title":"Anthropic open-sources FlashProve, a zero-knowledge prototype that proves a response came from committed model weights","deck":null,"takeaway":"On Nov 2, 2026 (X announcement Nov 3) Anthropic published FlashProve, an Apache-2.0 zkSNARK prototype that lets a verifier check that a model's output came from a specific committed set of weights without seeing them.","category":"research","category_label":"Research","importance":3,"confidence":"high","status":{"key":"confirmed","labels":["Confirmed"]},"sources":[{"n":1,"title":"Anthropic: Provable inference (FlashProve)","url":"https://www.anthropic.com/research/provable-inference","type":"official","group":"primary","domain":"anthropic.com"},{"n":2,"title":"FlashProve code (GitHub, Apache 2.0)","url":"https://github.com/anthropics/FlashProve","type":"code","group":"primary","domain":"github.com"},{"n":3,"title":"Anthropic on X (Nov 3)","url":"https://x.com/AnthropicAI/status/2117710942826123294","type":"official","group":"primary","domain":"x.com"},{"n":4,"title":"TechCrunch: Anthropic is working on a system to prove AI models' identity","url":"https://techcrunch.com/2026/11/03/anthropic-is-working-on-a-system-to-prove-ai-models-identity/","type":"press","group":"press","domain":"techcrunch.com"}],"official":3,"filed":"2026-11-04","updated":"2026-11-04","orgs":["Anthropic"],"title":"Anthropic open-sources FlashProve, a zero-knowledge prototype that proves a response came from committed model weights","summary":"On Nov 2, 2026 (X announcement Nov 3) Anthropic published FlashProve, an Apache-2.0 zkSNARK prototype that lets a verifier check that a model's output came from a specific committed set of weights without seeing them. Tested on Llama 4 Maverick, proving costs about 70–95x the GPU time of normal inference.","key_facts":["Model tested: Llama 4 Maverick (400B MoE, ~17B active), third-party 4-bit conversion","Proving ~110 tokens/s on one H100; proof ~19 MB per 1,024-token request; verification ~17 s on a 24-thread CPU","Overhead: roughly 70–95x the GPU time of serving the model on vLLM","Protocol properties (completeness, soundness, zero knowledge) proven in Lean; the fast C++ verifier is not formally verified","Limits: integer model under greedy decoding only, 1,024-token context, architecture not hidden, side channels out of scope","Anthropic's X post: 462k views (fxtwitter, Nov 4)"],"key_numbers":[],"tags":["safety","cryptography","zk-proofs","verification","open-source"],"science":null,"body_md":"## What happened\n\nAnthropic's research page describes FlashProve as \"an open-source performant prototype for provable inference.\" The provider publishes a\ncommitment to its weights once; afterwards it produces proofs for batches of requests, and anyone holding the commitment can check them\nwithout the weights. Anthropic's X post names two uses: making it harder to hijack a chat, and letting safety evaluators confirm that the\nmodel they tested is the one served to customers.\n\n## Why it matters\n\nEvaluators and governments have had no way to check that the deployed model is the audited one. The prototype is far from production cost\n(70–95x GPU time, short contexts, greedy decoding only), and Anthropic presents it as research, not a product.","disputed":[],"related":[],"people":[],"posts":[{"id":"x-anthropicai-2117710942826123294","title":"“New Anthropic research: progress on provable inference for large language models. Here…”","url":"https://postcutoff.com/p/x-anthropicai-2117710942826123294/"}],"videos":[],"models":[],"changes":[{"date":"2026-11-04","type":"filed","text":"Created"}],"provenance":{"agents":[{"model":"Claude Opus 5.5","maker":"Anthropic","tool":"Claude Code"}],"filed":"2026-11-04","run":null,"sources_read":null,"updated":"2026-11-04","human_review":null,"version":{"date":"2026-11-04"}},"gaps":[{"model_id":"gpt-6-astra","name":"GPT-6 Astra","cutoff":"2026-04","days_after":186,"in_training_data":false},{"model_id":"claude-opus-5-5","name":"Claude Opus 5.5","cutoff":"2026-06","days_after":125,"in_training_data":false},{"model_id":"gemini-4-argon","name":"Gemini 4 Argon","cutoff":"2026-03","days_after":216,"in_training_data":false},{"model_id":"grok-4-7","name":"Grok 4.7","cutoff":"2026-05","days_after":155,"in_training_data":false}],"short_url":null}