Grok 4.3 is positioned as a cheaper, stronger successor in xAI’s model line, with reported gains on instruction following and agentic customer support tasks. The main signal is benchmark positioning, but the writeup offers little technical detail beyond performance claims.
Grok 4.3 improves on cost-per-intelligence relative to Grok 4.20 0309 v2. It scores higher on the Intelligence Index while costing less to run the full benchmark suite. Grok 4.3 is one of the lowest-cost models at its intelligence level. It performs strongly on instruction following and agentic customer support tasks.