§ agents · storyline
GPT-5.5 and Gemini 3.1 Pro lack on-the-job learning, Skyfall AI finds
Skyfall AI's benchmark finds GPT-5.5 and Gemini 3.1 Pro lack on-the-job learning capabilities.
GPT-5.5 and Gemini 3.1 Pro Don't Learn on the Job — Skyfall AI's Benchmark Exposes the Hidden Ceiling on Agentic AI FourWeekMBA
§ sources1 publication · timeline below