This August 28, 2026 explainer covers the August 12, 2026 launch of Grok 4.6 by SpaceXAI (formerly xAI), organizing the background of the development that improved performance without increasing the scale of the base model, its position in major benchmarks, pricing, availability, and the corporate restructuring that sa The first half covers basic information and benchmarks drawn from the Cursor official blog's joint announcement, while the second half addresses features of Grok Build and past safety issues, providing context needed for implementation decisions. The piece frames Grok 4.6 as a model whose announcing company has changed

This benchmark deep-dive reports that Grok 4.6 scores 61 on the Artificial Analysis Intelligence Index, tying GPT-5.6 Sol and landing just behind Claude Opus 5 (63) and Claude Fable 5 (62), and notes that the model was launched with a focus on long-running agents and coding. Despite that framing, the benchmarks reveal The analysis highlights an Artificial Analysis measurement that xAI did not publish in its own announcement: a 65.7% non-hallucination rate, which matters more for customer-facing work than any coding score. On efficiency, Grok 4.6 reportedly finished long agentic tasks in about 53 turns compared to Claude Opus 5's 103

This analysis covers the August 12, 2026 Grok 4.6 release by SpaceXAI, reporting a score of 61 on the Artificial Analysis Intelligence Index, up 5 points from Grok 4.5 and 23 points from Grok 4.3, with pricing held flat at $2 per million input tokens and $6 per million output tokens, unchanged from the prior generation The piece provides critical cross-checking of vendor claims, noting that the "matches GPT-5.6 Sol" headline is accurate only on the nine-benchmark composite, while on xAI's own ten-row evaluation table Claude Fable 5 Max wins the most first-place rows. Grok 4.6 is reported as strongest on knowledge work and legal reaso

SpaceXAI (formerly xAI) released Grok 4.6 on August 12, 2026, with a stated focus on long-running agents, coding, and knowledge work, scoring 61 on the third-party Artificial Analysis Intelligence Index, tying GPT-5.6 Sol Max and trailing Claude Opus 5 and Claude Fable 5. The model improves five points over Grok 4.5 Hi For enterprises evaluating AI agents, the coverage highlights that Grok 4.6 is positioned for sustained multi-step task execution rather than single-shot scores, and the pricing strategy is described as deliberately competitive against both proprietary and open-source frontier models globally. The report includes a det