AI News Fastino Labs Open-Sources GLiGuard: A 300M Parameter Safety Moderation Model That Matches or Exceeds Accuracy of Models 23–90x Its Size May 13, 2026 0 34 FacebookXPinterestWhatsAppLinkedinReddItEmailPrintTumblrTelegramMix