OpenAI and Anthropic have both placed their coding CLI clients on GitHub, letting developers see the code that powers their AI‑assisted programming tools. A recent post by Bolin explains the core "agent loop" that coordinates user input, model responses, and tool calls. The loop repeatedly builds a prompt, sends it to the model, and either returns a final answer or executes a requested tool before querying the model again. The initial prompt is assembled from distinct components—system, developer, user, and assistant—each with a specific role, and includes instructions, tool definitions, sandbox settings, environment context, and the user’s message. Leia mais →

French AI startup Mistral AI introduced Devstral 2, a 123 billion‑parameter open‑weights coding model that achieved a 72.2 percent score on the SWE‑bench Verified benchmark. Alongside the model, Mistral released the Mistral Vibe command‑line interface, enabling developers to interact with Devstral models directly in their terminal, manage project context, and execute autonomous code changes. A smaller 24 billion‑parameter version, Devstral Small 2, scored 68 percent on the same benchmark and can run locally on consumer hardware. Both models support a 256,000‑token context window and are released under permissive open‑source licenses. Leia mais →

Google has broadened the reach of its AI coding assistant Jules by launching a command‑line interface and opening a public API. The new tools let developers invoke Jules directly from terminals, CI/CD pipelines, and chat platforms, reducing the need to switch between web pages and code repositories. Jules will continue to run on Google’s Gemini 2.5 Pro model and now supports more scoped, autonomous tasks while still prompting users when it encounters obstacles. The rollout follows recent feature upgrades such as memory, diff‑viewer enhancements, and broader version‑control support, positioning Jules as a deeper fixture in professional developer workflows. Leia mais →