← Retour au blog
tech 19 May 2026

The Last Six Months in LLMs in Five Minutes

Dive into the rapid advancements in LLMs over the past six months, from groundbreaking models to high-performing coding agents.

Article inspired by the original source
The last six months in LLMs in five minutes ↗ simonwillison.net

Introduction

The past six months have been a whirlwind for large language models (LLMs). If you feel like everything is changing at breakneck speed, you're not alone. These changes are not only rapid but also transformative for how we envision software development and automation. Let's dive into what has made headlines since November 2025.

The November 2025 Inflection Point

November 2025 marked a pivotal moment for LLMs, particularly in coding. The models deemed the "best" shifted hands five times in just a few weeks. Claude Sonnet 4.5, released on September 29th, was first overtaken by GPT-5.1, then Gemini 3, followed by GPT-5.1 Codex Max, before Claude Opus 4.5 reclaimed the crown. What really captured attention was the coding agents' ability to consistently produce high-quality code.

Coding Agents: A New Standard

OpenAI and Anthropic invested heavily in 2025 in reinforcement learning with verifiable rewards to enhance the quality of code generated by their models. By November, the results were remarkable. Coding agents moved from occasional reliability to consistent performance, allowing developers to use them daily without constantly fixing their errors.

The Rise of Ambitious Projects

With the improved LLMs, many developers took advantage of the holiday season to test these new tools. This led to a wave of innovative and sometimes audacious projects. One such project was the implementation of JavaScript in Python, known as micro-javascript.

Implications for the Future

Recent advancements suggest that we are only at the beginning of the LLM era in software development. The possibilities offered by these models extend well beyond simple coding. They also influence content generation, data analysis, and even strategic decision-making in companies.

Conclusion

The last six months have seen impressive advancements in the field of LLMs, changing the game for developers and decision-makers. These models are revolutionizing not only how we write code but also how we envision the future of automation and artificial intelligence.

Let's discuss your project in 15 minutes.

LLMs coding agents automation software development AI advancements
Deepthix newsletter · 100% AI · every Monday 8am

An AI agent reads tech for you.

Our AI agent scans ~200 sources per week and ships the best articles to your inbox Monday 8am. Free. One click to unsubscribe.

Visit the newsletter page →

Want to automate your operations?

Let's talk about your project in 15 minutes.

Book a call