Calculate Grok article costs, including the model-specific 20% Batch discount on Grok 4.3 and Grok 4.20.
Catalog rates verified . See verification and correction history.
Base text-token estimate, not an invoice. Excludes reasoning beyond the word estimate, output overshoot, cache writes/storage, tools, regional surcharges and editing. Unsupported tiers use Standard and say so. Check the calculation assumptions and budget for human editing.
Rate shown is for prompts under 200K tokens; $4.00 / $12.00 at or above that.
One 1,500-word article costs $0.00429 on Grok Build 0.1, xAI's cheapest tracked model, and $0.0125 on Grok 4.5, its most expensive. Grok Build 0.1 has no Batch tier. Availability and discounts differ by model; check the selector above.
This is a price ranking, not a writing-quality recommendation. Check the model-selection guide and real AI-written article examples before deciding what fits your editorial workflow.
One draft, a 250-word brief, 20% prompt overhead. Change any of it in the calculator.
| Model | 500w | 1,000w | 1,500w | 2,000w | 3,000w |
|---|---|---|---|---|---|
| Grok Build 0.1 | $0.00169 | $0.00299 | $0.00429 | $0.00559 | $0.00819 |
| Grok 4.3 | $0.00211 | $0.00374 | $0.00536 | $0.00699 | $0.0102 |
| Grok 4.20 Reasoning | $0.00211 | $0.00374 | $0.00536 | $0.00699 | $0.0102 |
| Grok 4.20 Non-Reasoning | $0.00211 | $0.00374 | $0.00536 | $0.00699 | $0.0102 |
| Grok 4.20 Multi-Agent | $0.00211 | $0.00374 | $0.00536 | $0.00699 | $0.0102 |
| Grok 4.7 | $0.00468 | $0.00858 | $0.0125 | $0.0164 | $0.0242 |
| Grok 4.6 | $0.00468 | $0.00858 | $0.0125 | $0.0164 | $0.0242 |
| Grok 4.5 | $0.00468 | $0.00858 | $0.0125 | $0.0164 | $0.0242 |
Cheapest first. These are list rates applied to a token model — see the methodology for exactly where that diverges from an invoice.
The 1,000w column is the per-1,000-words figure people quote: $0.00299 on Grok Build 0.1 at these settings.
Providers do not price the same way, and the differences change which model wins.
xAI's structure is genuinely different, and it matters for article writing. Output costs five times input on every Claude model and averages over six times across OpenAI and Gemini. Across the Grok range it averages about two and a half times and never exceeds three. Article generation is the most output-heavy workload there is — a 1,500-word draft is about 1,950 output tokens against a few hundred of input — so a narrow output premium works directly in its favour.
Grok 4.3 supports Batch at 20% off input, output and cache-read rates, and so do the three Grok 4.20 models (Reasoning, Non-Reasoning and Multi-Agent). Grok 4.5, Grok 4.6, Grok 4.7 and Grok Build 0.1 do not support Batch, according to their model pages. A missing discount is not the same as missing API support: check both for each model. xAI's separate 2× Priority service is not modeled here.
All eight Grok models change rate at 200K input tokens, doubling both input and output. The published window goes to 500K or 1M, so the advertised context and the single-rate context are not the same number.
Compare the 20% Grok 4.3 and 4.20 Batch discount with the supported 50% discounts elsewhere, not with competitors' standard prices. Tool invocations and reasoning output can add costs beyond this word-based estimate; a low output/input ratio is not a quality test.
Five other models wrote the same commissions under the same rules — unedited, blind, with recorded usage repriced at current rates. Read them to calibrate what a raw API draft looks like at any price before you budget on one.
USD per million tokens, cheapest output first.
| Model | Input | Cached in | Output | Context |
|---|---|---|---|---|
| Grok Build 0.1 Rate shown is for prompts under 200K tokens; $2.00 / $4.00 at or above that. | $1.00 | $0.200 | $2.00 | 256K |
| Grok 4.3 Rate shown is for prompts under 200K tokens; $2.50 / $5.00 at or above that. | $1.25 | $0.200 | $2.50 | 1,000K |
| Grok 4.20 Reasoning Rate shown is for prompts under 200K tokens; $2.50 / $5.00 at or above that. | $1.25 | $0.200 | $2.50 | 1,000K |
| Grok 4.20 Non-Reasoning Rate shown is for prompts under 200K tokens; $2.50 / $5.00 at or above that. | $1.25 | $0.200 | $2.50 | 1,000K |
| Grok 4.20 Multi-Agent Rate shown is for prompts under 200K tokens; $2.50 / $5.00 at or above that. | $1.25 | $0.200 | $2.50 | 1,000K |
| Grok 4.7 Rate shown is for prompts under 200K tokens; $4.00 / $12.00 at or above that. | $2.00 | $0.500 | $6.00 | 500K |
| Grok 4.6 Rate shown is for prompts under 200K tokens; $4.00 / $12.00 at or above that. | $2.00 | $0.500 | $6.00 | 500K |
| Grok 4.5 Rate shown is for prompts under 200K tokens; $4.00 / $12.00 at or above that. | $2.00 | $0.300 | $6.00 | 500K |
Source: xAI's published pricing. Compare against the other providers on the full pricing table.
Answers specific to this provider's rate structure.
Grok Build 0.1 is the cheapest tracked xAI model for a full draft. All eight Grok models sit in a narrow band compared with the ranges the other providers publish.
Yes: Grok 4.3 and the three Grok 4.20 models reduce input, output and cache-read rates by 20% on Batch. The other tracked Grok models do not support Batch. The calculator checks model support and uses the actual discount rather than assuming 50% everywhere.
Because of the output premium. Output costs five times input on every Claude model and averages over six times across OpenAI and Gemini. Across the Grok range it averages about two and a half times and never exceeds three. Article generation is the most output-heavy workload there is, so a narrow premium works directly in its favour.
All four models double both input and output rates at or above 200K input tokens. The published windows go to 500K or 1M, so the advertised context and the single-rate context differ. An article workflow never approaches it.
Compare against batch rates, not standard ones. Grok’s narrow output premium is a real advantage at standard rates, but a competitor’s 50% batch discount can close it entirely if your work can run asynchronously.
Price your workflow here, then check it against the other providers before committing. At article volumes the difference between two reasonable models is usually smaller than a single hour of editing. When the work can run asynchronously, compare Grok's batch rate — not its standard rate — against DeepSeek, where no Batch tier is published. DeepSeek instead offers a separate off-peak discount.