Grok 4

Reviewers say it shows strong performance on coding benchmarks and competitive results against leading models. However, reviewers disagree on whether those benchmark scores match how the model actually feels when you use it day-to-day for coding work.

Reviewers note that benchmark improvements don't always translate to real-world coding feel, and costs rise noticeably as capability increases. The model also lacks true general reasoning and common sense outside controlled test conditions.

Reviewers are split. It sits below top-tier closed models in price while remaining capable, but reviewers disagree on whether current pricing is good efficiency or a step backward. It's more expensive than many open-weight alternatives.

Reviewers recommend it for developers and technical users who want a fast, capable reasoning and coding model at a competitive price point. It's not the right fit if you need the single best coding performance regardless of cost, or if you want a model with proven real-world reliability over raw benchmark scores.

Reviewers note that the model shows a fast iteration pace with clear generation-over-generation improvement on reasoning and coding benchmarks.