Grok 4.5 Launches Across All Platforms: Aggressive Pricing and Strong Coding Benchmarks

eyishazyer · x · 2026-07-24

xAI announced that Grok 4.5 is live across all its platforms simultaneously on July 22, abandoning staggered rollouts.

Coding Performance: On VulcanBench v3, which tests multi-file repo work, the model solved 21 out of 23 tasks with an overall score of 91.3%, beating Fable 5 and GPT-5.6 Sol. While it doesn't sweep all benchmarks, it wins on specific, real-world tests like SWE Marathon.

Pricing: Priced aggressively at $2 per million input tokens and $6 for output, significantly cheaper than competing flagship models.

Distribution: Beyond the model itself, the same-day launch across xAI's entire ecosystem is a distribution move that Anthropic and OpenAI haven't matched yet.

Related event: xAI's Flagship Grok 4.5 Launches Across All Platforms Simultaneously(16 posts)→

Original post →

More from Companies & People

Companies & People channel →