Get the app →
BriefTea logoBriefTea

Every story in sixty words

Get the free app no ads · no paywall

Technology · 6 October 2026

Mistral launches new AI model, pitching open weights for security

Mistral launches new AI model, pitching open weights for security

Original: Mistral AI Vector: Pantarch via Wikimedia Commons (CC BY-SA 4.0)

French AI firm Mistral has launched a public preview of its new Large 4 model, nicknamed Le Chonk, with its full 'open weights' due on October 27. The company says open weights protect users from a vendor suddenly turning off or limiting a model's capabilities. Mistral reports Large 4 scored highest on a cyber test, while some closed models scored near zero.

Reported by implicator.ai · How we write briefs

Get the app

Understand this story

How can a vendor limit an AI model? Why would AI firms share their code? What does a cyber test measure?

More briefs

WorldIntel began new IT job cuts on 16 November, affecting hundreds of staff GeneralKitchen porter jailed for at least 22 years for killing former Microsoft executive GeneralAI could displace 19% of UK knowledge workers, a new report warns GeneralWoman, 27, shot in bed as Police investigate attempted murder BusinessUber-style drivers and other contractors lose protections under new US Labor Department rule WorldUS Treasury chief Bessent says Nvidia chip curbs on China are enough Browse all stories →