Inside Grok 4.5: How Cursor Collaboration and Real Developer Data Shaped the Model
eyishazyer · x · 2026-07-30
The thread delves into the training details of Grok 4.5. This marks the first time xAI collaborated with Cursor, training the model on real developer-agent sessions rather than relying solely on synthetic benchmarks.
Notably, Cursor's own writeup honestly admitted that Grok 4.5's benchmark edge might partly come from an old Cursor codebase snapshot leaking into the training data. The author points out that Grok has previously suffered from a disconnect between top-tier benchmark scores and real-world user rankings (e.g., Grok 4 ranked around #66 globally), and it remains to be seen if 4.5 breaks this pattern.
More from Companies & People
- AI Lab Professor: 3 Unwritten Rules for Writing a Strong PhD SOP — prof_kamilov · 2026-07-30
- Searching for Gemini on Edge? Microsoft Slaps a Copilot Ad on Top — Shadow_Fury69 · 2026-07-30
- Anthropic Clashes with Senate Over AI Safety Bill, Criticized for Wanting 'Check-the-Box' Compliance — neil_chilson · 2026-07-30
- Chinese Robot Makers Overtake Foreign Rivals with 57% Domestic Market Share — PeterDiamandis · 2026-07-30
- AI-Powered Founders School Guarantees Teens $1M Profit or Tuition Refund — nateliason · 2026-07-30
- Microsoft AI CEO Bets on Cheap Specialist Models Over General-Purpose Frontier — The Decoder · 2026-07-30