Google Enhances Gemini CLI Capabilities

Google has unveiled two new features that enhance the capabilities of Gemini CLI. The Gemini 2.5 Computer Use model allows for the development of AI agents capable of manipulating user interfaces, outperforming competitors in web and mobile control benchmarks. This model is available via the Gemini API.

Additionally, Google has introduced support for extensions in Gemini CLI, enabling connections with various tools and personalization of developer workflows. A list of launch partners, including Figma, Shopify, and Stripe, has been shared.

OpenAI has announced an 'app platform' within ChatGPT that allows third-party apps to run directly inside ChatGPT. Partnerships with Spotify and Canva have been announced, and an Apps SDK is available for other third-party app providers to join the ecosystem.

OpenAI has also added new AI models GPT-5 Pro and Sora 2 to its APIs and introduced the GPT-realtime-mini speech-to-speech API, a fast low-latency model maintaining high quality for voice agents and conversational applications.

Figure AI has unveiled its third-generation humanoid robot, Figure03, powered by the Helix AI system. The robot features a 5-hour battery life and wireless charging, with mass production planned at a factory capable of producing 12,000 units annually.