Grok 4.3

Grok 4.3 is a language model built by xAI for developers and business workloads. It runs at approximately 100 tokens per second, offers a one-million-token context window, and reasoning is built in by default, billed at the same rate as regular output tokens. Knowledge cutoff: December 2025.

Pricing is $1.25 per million input tokens and $2.50 per million output tokens — approximately 40–60% cheaper than the previous Grok 4.20. A full Artificial Analysis benchmark run costs $395 (vs $3,959 for GPT-5.5 and $4,811 for Claude Opus 4.7). The model has built-in tool calling: web search, X search, Python execution, RAG, and Excel/PDF/PowerPoint generation.

On the Artificial Analysis Intelligence Index, Grok 4.3 scored 53 points (8th place), below GPT-5.5 (60 pts), Claude Opus 4.7 and Gemini 3.1 Pro (both 57 pts). The model ranks first on Val's AI CaseLaw and CorpFin benchmarks but 13th on hard coding and math. Andon Labs reported a "narcolepsy" problem in autonomous tests. Available via OpenRouter, xAI API, and the Hermes agent (Nous Research).