xAI: Grok 4.3

Grok 4.3 is a text-and-image model from xAI with a 1-million-token context window, tool use, and built-in reasoning support. Structured output availability is not confirmed in current documentation. The large context window makes it a practical option for tasks involving lengthy documents or multi-turn conversations, and tool support extends its usefulness to agentic workflows. At $1.25 per million input tokens and $2.50 per million output tokens, it sits in a budget-friendly tier relative to many reasoning-capable models. Its blended benchmark score of 57.7 covers only three benchmarks, so treat that figure as partial rather than comprehensive. Coding tasks score notably higher (69.7) than agentic tasks (39.8), which suggests Grok 4.3 is a stronger fit for code-assistance use cases than for complex multi-step agent pipelines. Teams prioritizing cost efficiency alongside coding support have reasonable grounds to shortlist it, though the limited benchmark coverage warrants caution before committing to production use.