OpenAI, Google, and Anthropic have all announced new models, so here’s a quick breakdown of what they’re good for as I understand it. GPT-6.1 Terra (OpenAI) - Built for computer use and visual tasks - Good for agents that interact with websites and apps - Worth checking out if you’re building browser automation or computer-use agents Gemini 4 Carbon (Google) - Focused on coding and cybersecurity - Good for complex engineering tasks and finding vulnerabilities - Still limited in availability, but interesting to keep an eye on Claude Haiku 5.5 (Anthropic) - Built for speed and lower-cost tasks - Good for summarization, classification, and subagents - Could be a solid choice for workflows that make a lot of API calls One thing worth remembering is that you don’t always need the most powerful model. Sometimes a smaller, cheaper model gets the job done just as well. Also, if you’ve been testing any of these, drop your experience in the comments.