https://lmarena.ai/leaderboard/ The leaderboard at lmarena.ai/leaderboard offers a comparative overview of leading AI models across several domains, including text generation, web development, vision (image tasks), search, copilot/code, text-to-image, and image editing. Models are ranked based on their performance scores, with dedicated tabs for deeper domain-specific insights. -------------------------------- Use it to decide which LLM tool best suits your application. Good to see some Open Source models right up there with the leading closed models, eg. Qwen3, deepseek, kimi, glm and mistral

https://www.perplexity.ai/comet Comet Browser is a new AI-powered web browser developed by Perplexity, designed to enhance productivity and user experience by integrating advanced artificial intelligence tools directly into the browsing environment. Built on Chromium, Comet supports familiar features like Chrome extensions and bookmarks, making the transition seamless for most users. Comet is positioned as a next-generation browser that transforms passive web surfing into an interactive, AI-driven workflow, aiming to set a new standard for intelligent, agentic browsing. ----------------------------------------------------------- I recommend trying it out. Its quickly become my preferred browser so I don't need to have multiple tabs open to use multiple AI tools.

https://huggingface.co/ Hugging Face (huggingface.co) is a leading AI platform and open-source community where developers, researchers, and organizations collaborate to create, share, and deploy machine learning models, datasets, and applications, especially in the field of natural language processing (NLP). The easiest ways to run open source AI models locally are with user-friendly tools like https://ollama.com/ and https://lmstudio.ai/, which remove the need for complex setup or coding skills and work well on Windows, Mac, and Linux systems. Just keep in mind that the size (GB) of your AI model should ideally fit fully within your GPU VRAM, otherwise it will offload to the CPU and drastically reduce the speed (tokens per second).