SpaceXAI has launched Grok 4.5, its newest AI model built for coding, agentic tasks, and knowledge work.

Announcing Grok 4.5, our first model trained specifically for coding and agents. It was trained with Cursor and offers frontier intelligence at leading speeds and cost efficiency.https://t.co/i8HpU7w64k pic.twitter.com/oBjGtTsoNc

— SpaceXAI (@SpaceXAI) July 8, 2026

The model is available today in Grok Build, Cursor, and the SpaceXAI console, according to the company. Grok 4.5 is not yet available in the EU, with availability expected in mid July.

The launch marks SpaceXAI’s first major model release since acquiring Cursor, the AI coding startup behind one of the most widely used developer tools in the market.

SpaceXAI is positioning Grok 4.5 around practical engineering work rather than general chatbot use. The company says the model was trained across coding, science, engineering, and math datasets, with reinforcement learning focused on long running software engineering and technical tasks.

In benchmark data shared with the launch, SpaceXAI said Grok 4.5 reached 62% on DeepSWE, 83.3% on Terminal Bench 2.1, and 64.7% on SWE Bench Pro. The company also said the model runs at 80 tokens per second and uses fewer output tokens than comparable models on software engineering tasks.

The pricing is also part of the pitch. Grok 4.5 costs $2 per million input tokens and $6 per million output tokens, which SpaceXAI argues gives the model a stronger cost profile for enterprise workloads.

The release gives SpaceXAI a more direct entry into the market for coding agents and workplace AI tools. It also tightens the link between Grok and Cursor, turning the model into a product aimed at developers, analysts, and companies trying to automate technical work across code, spreadsheets, documents, and presentations.