In a nutshell: In November 2025, a turning point in LLM development was reached: coding agents became production-ready, while competition for the best model intensified. Since then, the community has enthusiastically experimented with new possibilities and driven innovative infrastructure projects forward.
A review of the most important developments in the AI world since November 2025: the race for the best model intensified dramatically, coding agents became production-ready, and a new infrastructure movement emerged from experimental passion.
The author summarized the rapid developments of the last six months in a lightning talk at PyCon US 2026, identifying November 2025 as a critical turning point in the AI industry.
During this period, the recognized best model switched five times between the three major providers: Claude Sonnet 4.5 was displaced by GPT-5.1, followed by Gemini 3, GPT-5.1 Codex Max, and finally Claude Opus 4.5. What was particularly notable was not only this constant shift at the top, but the underlying technological shift.
The main event was the maturation of coding agents. After OpenAI and Anthropic spent most of 2025 leveraging Reinforcement Learning from Verifiable Rewards to improve their models’ code quality, the fruits of this work became evident in November. Coding agents crossed a critical quality threshold: they now functioned reliably enough to be deployed as daily tools without developers spending most of their time on debugging.
During the winter break from December to January, the community extensively experimented with these new possibilities. A particularly interesting project was a vibe-coded JavaScript implementation in Python called micro-javascript – a loose port of MicroQuickJS that traverses multiple layers of complexity: JavaScript code in Python, executed in Pyodide, which runs in WebAssembly, which in turn is executed in JavaScript in the browser.
The enthusiasm of this phase also led to the launch of the Warelay project, an infrastructure initiative that emerged from this phase of experimental innovation and has since continued to grow.