Grok 4.1 Fast Pricing

Grok 4.1 Fast API pricing

Grok 4.1 Fast API pricing is $0.2 per million input tokens and $0.5 per million output tokens, verified September 14, 2026.

Input (per 1M tokens)

$0.2

Output (per 1M tokens)

$0.5

Cached input (per 1M tokens)

$0.05

Context window

2,000,000 tokens

Max output

16,384 tokens

Capabilities

Grok 4.1 Fast supports image input, tool use, long context.

Grok 4.1 Fast API pricing has not changed since June 28, 2026.

Price-stable since June 28, 2026.

Frequently asked questions

Short answers to the most common questions about Grok 4.1 Fast pricing, limits, and capabilities.

Notes

Fast variant optimized for high throughput. Among the cheapest frontier-adjacent options on the market at $0.20/M input. Cached input drops to $0.05/M with prompt caching enabled. [2026-07-12] Context Window updated 1,000,000 → 2,000,000 per multiple corroborating sources (x.ai news, datastudios.org, Oracle OCI docs, mindstudio.ai). Prices unchanged ($0.20/$0.50, cache read $0.05). CAUTION: one source reports an 8,000-token per-response cap vs 16,384 Max Output in table — unresolved, recheck next run. [2026-07-13] Prices re-verified unchanged ($0.20/$0.50, cache read $0.05) via mem0.ai + costgoat. Config conflict PERSISTS and widened: datastudios.org claims 128K context / 8,192 max output; aimlapi.com docs claim 2M context / 16K playground cap. Left 2M / 16,384 in place — needs manual check against docs.x.ai; recheck next run. [2026-07-27] Verification attempt produced no clear result — recheck next run. Search reports Grok 4.1 Fast is no longer listed on xAI's current public pricing page (possibly legacy/enterprise-only); no fresh price confirmation found. Prices untouched. Config conflict from prior runs still unresolved. Deprecation candidate — Khaled to decide. [2026-07-29] Prices re-verified unchanged ($0.20/$0.50) via developer.puter.com + benchlm.ai — first price confirmation since the 07-27 delisting concern. Model still not prominent on xAI's main pricing page; context/max-output config conflict from prior runs still unresolved (needs manual docs.x.ai check). Deprecation decision still with Khaled. [2026-08-17] Prices re-verified UNCHANGED ($0.20 input / $0.50 output) via pricepertoken (SKU xai-grok-4.1-fast), aifreeapi.com and costgoat. CONTEXT CONFLICT RESOLVED: sources now consistently state a 2M-token context window, matching the stored value — the 128K/8K figures from the July runs are treated as stale. Max Output 16,384 still unconfirmed but no longer disputed. Model remains absent from xAI's headline pricing page (4.5/4.6 occupy it); deprecation decision still with Khaled. [2026-09-07] DELISTING QUESTION NOW RESOLVED — THIS MODEL IS RETIRED. Two vendor-grade sources agree: xAI's own migration page (docs.x.ai/developers/migration/may-15-retirement) and Oracle OCI's model docs both state Grok 4.1 Fast was deprecated 2026-05-15 and RETIRED 2026-08-15. Per xAI, requests to the retired slug no longer run this model — they silently redirect to grok-4.3 and bill at grok-4.3 rates ($1.25/M input, $2.50/M output), i.e. 6.25x the input price and 5x the output price this row currently advertises. The retirement date passed three weeks ago. PRICE FIELDS DELIBERATELY LEFT UNTOUCHED: rewriting them to grok-4.3's rates would misrepresent a dead SKU as a live one, and changing Status is outside this task's remit. ACTION FOR KHALED — HIGHEST PRIORITY THIS RUN: set Status to Retired (or Deprecated). While the row reads Live at $0.20/$0.50, the calculator is quoting users a price for a model that cannot be called and would silently bill them ~6x more. This is the one item this week that is actively wrong in a user-facing way.