Skip to content
The Nexus
DossierENTITY

software engineering benchmarks

Coverage of software engineering benchmarks in the Nexus archive.

Earliest in view: Jun 2 · 18:12 UTCMost recent: Jun 2 · 18:12 UTC
Co-mentioned in this coverage
Microsoftorganization1MAI-Thinking-1topic1OpenAIorganization1
Recent coverage
  • TECHNOLOGYJun 2 · 18:12 UTCTHE VERGE
    Microsoft’s first advanced reasoning AI is here

    Microsoft announced its first in-house advanced AI model, MAI-Thinking-1, at Build 2026. The model is described as a medium-sized system that matches leading models on key software engineering benchmarks and was trained from scratch without relying on third-party models like those from OpenAI. The company has shifted toward developing its own AI models after renegotiating its partnership with OpenAI.