xAI Retires 8 Older Grok Models, Consolidates to Grok 4.3

xAI announced the retirement of eight older Grok models effective May 15, 2026 at 12:00 PM PT. The deprecated models include grok-4-1-fast-reasoning, grok-4-1-fast-non-reasoning, grok-4-fast-reasoning, grok-4-fast-non-reasoning, grok-4-0709, grok-code-fast-1, grok-3, and grok-imagine-image-pro. All requests to these model slugs will automatically redirect to Grok 4.3 with preset reasoning efforts (low for reasoning models, none for non-reasoning models). Grok 4.3 is xAI's flagship model, supporting a 1 million token context window, four reasoning effort levels (none, low, medium, high), and top leaderboard scores in agentic tool calling and instruction following. Its pricing is $1.25 per million input tokens and $2.50 per million output tokens, which may differ from the rates of the retired models.

To avoid unexpected billing changes, xAI strongly recommends that developers explicitly update their API requests to use Grok 4.3 with the appropriate reasoning effort before the retirement date. For code workloads previously using grok-code-fast-1, Grok 4.3 delivers improved agentic coding and web dev capabilities. For image generation, grok-imagine-image-pro redirects to grok-imagine-image. xAI has provided a migration guide and recommends leveraging Grok 4.3's new reasoning effort levels—low for general use, medium or high for deeper reasoning—to fine-tune performance and cost. This consolidation simplifies xAI's model lineup and ensures all users benefit from the latest advancements.