Anthropic recently acquired the startup Vercept to bolster its Claude models, which now feature advanced "computer use" and coding capabilities through tools like Claude Code and Opus 4.6. Meanwhile, OpenAI has launched the GPT-5.2 and 5.3 series, including the high-speed Spark model and specialized Codex versions for developers and healthcare providers. While Anthropic faces scrutiny for loosening safety protocols amidst military partnerships and government pressure, researchers are increasingly comparing these frontier models for their effectiveness in autonomous engineering and creative tasks. Collectively, the texts illustrate a shift toward agentic AI that can independently plan, execute, and debug complex workflows across various professional sectors.