Grok 4.5 arrives with 500K tokens and aggressive pricing: the AI war is now about cost, code, and agents Grok 4.5 arrives with a 500K-token context window and aggressive pricing for code and agents xAI has launched Grok 4.5 as its new flagship model for code, agents, and technical tasks. According to xAI’s documentation, Grok 4.5 is designed for “code and everything else” and operates with a 500K-token context window. The model also supports function calling, structured outputs, configurable reasoning, and both text and image input. The announced API pricing is US$2 per 1 million input tokens and US$6 per 1 million output tokens. Axios reported that Grok 4.5 was launched primarily as a tool for coding and agentic work, not just as a standard chatbot. The report also notes that the launch comes after SpaceXAI acquired Cursor, one of the most widely used tools among AI developers. Elon Musk described Grok 4.5 as an “Opus-class” model, indicating that xAI wants to position it in the same tier as advanced models built for complex tasks. In video-based tests, the model was used to create 3D games, interactive scenes, SVGs, and systems involving physics, audio, and interaction logic. In one example, a 3D city was built using 1.35 million tokens for about US$8, with Fable 5 generating the specification and Grok 4.5 executing the code. The launch highlights an important shift in the AI market: competition is no longer only about who has the smartest model, but about who delivers the best balance of cost, speed, context, and practical execution. With a 500K-token context window, US$2/US$6 per million tokens pricing, and a strong focus on programming, Grok 4.5 enters the market as a powerful alternative for agents, automation, and software development at scale. https://codingfleet.com/blog/swe-bench-pro-leaderboard-2026/