Elon Musk to Release Grok 4 and Grok 4 Code AI Models After July 4 With Advanced Reasoning and High Benchmark Scores

Elon Musk and his xAI team are preparing to release Grok 4, the latest iteration of their AI language model, shortly after July 4, 2025. Grok 4 is expected to be a major upgrade over its predecessor, Grok 3.5, with advanced reasoning capabilities and a mission to rewrite and correct the entire corpus of human knowledge by fixing errors, filling gaps, and retraining on improved data. The release will include at least two models: Grok 4, an all-rounder in language, math, and reasoning, and Grok 4 Code, a developer-focused model optimized for coding tasks. Leaked benchmark results suggest Grok 4 significantly outperforms previous models, achieving scores such as 35% on the Humanities Last Exam (HLE) benchmark and up to 45% with reasoning, 87-88% on the GPQA benchmark, and 72-75% on the SWE Bench for Grok 4 Code.

These benchmarks indicate Grok 4 could become a state-of-the-art language model, with Elon Musk personally involved in finalizing the model through a last major training run. The AI’s enhanced accuracy and reasoning aim to address the limitations of prior models trained on flawed or biased data. The Grok 4 launch is highly anticipated as it promises to advance AI capabilities, particularly in knowledge accuracy and coding specialization.