xAI prices Grok‑4.3 API

xAI has posted public pricing for Grok 4.3 in its developer documentation, giving software teams a current reference point for using the company’s flagship model through the API. The docs list Grok 4.3 at $1.25 per 1 million input tokens, $0.20 per 1 million cached input tokens and $2.50 per 1 million output tokens. The same model page says Grok 4.3 has a 1 million-token context window and supports configurable reasoning settings. The documentation update lands as xAI consolidates older text models into Grok 4.3. xAI’s migration guide says several earlier model slugs were retired on May 15, 2026, and requests to those slugs now redirect to Grok 4.3, with billing at standard Grok 4.3 rates. That gives developers a concrete cost baseline as they decide whether to keep legacy integrations in place or switch models explicitly. (docs.x.ai) ### What exactly did xAI publish for Grok 4.3? xAI’s Grok 4.3 model page lists the model as the company’s “most advanced flagship model” and sets out the core commercial terms in one place. The page shows text-and-image input, a 1 million-token context window, function calling, structured outputs and four reasoning levels: none, low, medium and high. (docs.x.ai) The pricing table on xAI’s broader pricing page repeats the same numbers. It lists Grok 4.3 at $1.25 per 1 million input tokens, $0.20 per 1 million cached input tokens and $2.50 per 1 million output tokens, all priced in U.S. dollars. ### How does the pricing fit into xAI’s current model lineup? xAI’s models page says Grok 4.3 is now the default recommendation “for everything else” outside dedicated audio, image and video APIs. (docs.x.ai) The same page says older model aliases can point users to the latest stable or latest available version, depending on the naming convention they choose. May 15, 2026 is the key date in the transition. xAI’s migration guide says retired slugs including grok-4-fast, grok-4-1-fast, grok-4-0709, grok-code-fast-1 and grok-3 now resolve to Grok 4.3, with reasoning or non-reasoning behavior mapped to low or none reasoning effort by default. (docs.x.ai) ### What other charges show up around the base token price? (docs.x.ai) xAI’s pricing page says tool use adds separate charges on top of token billing. Web Search, X Search and Code Execution are each priced at $5 per 1,000 calls, while attachment search is $10 per 1,000 calls and collections search is $2.50 per 1,000 calls. The same page says total costs can rise with query complexity because the agent decides how many server-side tools to invoke. xAI also says token billing can include input, reasoning, completion, image and cached prompt tokens, depending on the request. (docs.x.ai) ### What does xAI say developers should do with older integrations? xAI’s migration guide says developers do not need to change code immediately to avoid breakage because retired slugs continue to resolve. (docs.x.ai) The company also says continuing to use deprecated slugs after May 15 can produce different bills because those requests are now charged at Grok 4.3 rates rather than the original model rates. The guidance recommends changing the “model” field explicitly in API requests. xAI says that gives users control over the reasoning level they pay for instead of accepting the redirect defaults. ### Where can teams check the current numbers? xAI’s developer docs now carry the main reference pages. The Grok 4.3 model page lists the model capabilities, rate limits and token prices, while the pricing page lists chat, voice, image, video and tool-use charges in one table. (docs.x.ai) May 12, 2026 is the last-updated date shown on the models page, and the May 15, 2026 migration guide sets out the redirect and billing changes for retired model slugs. (docs.x.ai) Those pages are the current public record for Grok 4.3 API pricing and the next step for developers reviewing existing deployments. (docs.x.ai 1) (docs.x.ai 2)