Evaluator Bench
Cloud Security Alliance (research note), 2026-05-05; retrieved 2026-09-15 tier 4 press (primary)confirmed

CAISI Frontier Testing Agreements Reach Five Labs

https://labs.cloudsecurityalliance.org/research/csa-research-note-caisi-frontier-ai-testing-agreements-20260/

Google DeepMind, Microsoft and xAI signed early-access agreements on 5 May 2026; OpenAI and Anthropic since September 2025; more than 40 evaluations completed; access includes versions with guardrails stripped back; DeepSeek V4 Pro evaluated without developer cooperation.

Signals citing this source

  • US CAISI (NIST), X (for): No commercial products.
  • US CAISI (NIST), A (for): Five labs now give CAISI pre-deployment access (OpenAI and Anthropic since September 2025; Google DeepMind, Microsoft and xAI from 5 May 2026), including versions with guardrails stripped back; more than 40 evaluations completed, one of a foreign model without developer cooperation.“provide CAISI with model access that includes versions with safety guardrails stripped back”
  • US CAISI (NIST), S (against): Participation is voluntary and agreements were jointly scoped with the two founding labs; the center can only assess models developers choose to share.“CAISI’s authority to conduct these evaluations rests entirely on voluntary participation”

Ledger rows citing this URL

  • none