AINewsnow

19 MB proves a 400B model reply at 72x cost

19 MB proves a 400B model reply at 72x cost Anthropic's FlashProve proves a reply came from a specific 400-billion-parameter model — without showing anyone the weights. The headline numbers are 110 tokens per second, a 19 MB proof, and a 70–95x GPU overhead. I cloned the repo and audited what those…

Read the full story at DEV Community — Machine Learning ↗

Timeline · 1 report

  1. 2026-11-05 00:34 · DEV Community — Machine Learning
    19 MB proves a 400B model reply at 72x cost

More stories

  1. Anthropic releases FlashProve, a zero-knowledge tool to verify which AI model answered a query — Runway Research
  2. Anthropic product chief Cat Wu departs, joins OpenAI to lead product management — Business Insider AI
  3. Investigating vault particles with Claude — Anthropic News
  4. Anthropic reports 1,100% revenue surge and $895B compute commitments ahead of IPO — Reuters AI
  5. TeraWulf secures Google credit backstop for Anthropic data-center project in Kentucky — Reuters AI
  6. Anthropic cuts Claude Code subagent usage limit consumption — Anthropic (X)
  7. The Briefing: Work at the Frontier Livestream — Anthropic News
  8. Investor Guides Highlight AI Stocks to Buy in November — The Globe and Mail

Get the daily brief of stories like this at 6:30 every morning →