Section

AI

Data stories on AI models, compute and adoption.

AI18 November 2026

DaisyRose-27B scores 12% on Terminal-Bench 4.1

RadixArk released DaisyRose-27B and reported a 12.0% score on Terminal-Bench 4.1, up from 5.6% for the base model.

AI17 November 2026

GPT-6.2 Astra scores 62.5% on repeat financial-analysis tests

OpenAI released GPT-6.2 Astra and Sol, and Astra answered 62.5% of the benchmark questions correctly on every attempt at its two highest tested settings.