Grok 4.5

• 🆕 xAI's coding-focused model for agentic software engineering, released 2026.
• 🧠 Reasoning, vision, function calling, structured outputs, web search built in.
• 📏 500K-token context window for large codebases and workflows.
• 🔧 Trained alongside Cursor on coding, science, engineering, math data.
• 🎯 Handles debugging, refactoring, and end-to-end app building.
• 🏢 Available via xAI console, Grok Build, and Cursor.
• 📚 Self-reported 83.3% on Terminal Bench 2.1.

xAI

xAI is an American artificial intelligence company and wholly owned subsidiary of SpaceX. The company develops AI systems under the Grok brand, spanning language models, image generation, and video synthesis. xAI has quickly established itself as a multimodal…

Grok 4.5 is xAI's coding-oriented flagship, positioned for agentic software engineering, workflow automation, and knowledge work. xAI describes it as trained on datasets spanning coding, science, engineering, and math, and notably co-developed with Cursor to learn from real development environments. It supports vision, function calling, structured outputs, and web search, with a 500K-token context window for large engineering tasks.

Compared with its same-family predecessor Grok 4.3, Grok 4.5 shifts emphasis toward agentic coding and end-to-end app building, from challenging Rust and C/C++ tasks through production. Where Grok 4.3 was positioned as a general-purpose flagship, Grok 4.5's strengths lie in reasoning quality and engineering tasks.

On xAI's own reported benchmarks, Grok 4.5 scores 62.0% on DeepSWE 1.0, 83.3% on Terminal Bench 2.1, and 64.7% on SWE Bench Pro resolve rate. These figures are vendor-reported and measured within provider-specific harnesses.

The model also succeeds earlier Grok generations such as Grok 4.20 and its multi-agent variant, and complements xAI's separate agentic coding tool Grok Build.