Grok 4.5 Is Now on Chat Smith

Chat Smith has added Grok 4.5 to its model lineup — xAI's new flagship built for coding, agentic workflows, and knowledge-intensive tasks, available now for all Chat Smith Pro users at no extra cost.

xAI launched Grok 4.5 publicly on July 8, 2026, co-trained with Cursor on trillions of tokens drawn from real developer activity. It is a mixture-of-experts model positioned as xAI's strongest release to date, with a focus on coding performance, agentic task resolution, and long-context reasoning.

Strong coding and agentic benchmarks. Grok 4.5 scores 64.7% on SWE-Bench Pro and 83.3% on Terminal-Bench 2.1. On SWE Marathon — a benchmark measuring long-horizon agentic task completion — it leads with 29.0%, ahead of comparable frontier models.

Fast throughput with a large context window. The model is served at 80 tokens per second across a 500K-token context window. Independent measurements from Artificial Analysis clocked it at approximately 91–120 tokens per second through the API.

High token efficiency for agent workflows. On SWE-Bench Pro tasks, Grok 4.5 uses an average of 15,954 output tokens per task — roughly 4.2 times fewer than comparable frontier models at the same tasks. For teams running AI agents at volume, this translates directly to lower costs and faster completions.

Broad knowledge-work capability. Beyond software engineering, Grok 4.5 handles long documents, multi-step research tasks, and structured knowledge work. Artificial Analysis places it at #4 on its Intelligence Index (score: 54) out of 187 tracked models, and it ranks first for agentic tool use.

Grok 4.5 is available immediately to all Chat Smith Pro subscribers. No upgrade, no additional charge — open the model picker and select Grok 4.5 to get started.

Chat Smith's goal is to bring the best available AI models into one place, so users can choose the right tool for each task without switching platforms.

Grok 4.5 is available now for all Chat Smith Pro users at https://chatsmith.io/model/grok-4-5.