Grok 4 free? The short answer
A $0 consumer plan with limited AI usage is listed, but the available information does not confirm that the original Grok 4 model is currently selectable for free. Consumer access is separate from API billing. The original grok-4-0709 API model ID was retired and requests to it redirect to Grok 4.3; the cited retirement notice does not specify the billing treatment for redirected requests. (xAI plans; model retirement notice)
“Grok 4” may refer to the original model, a later model in the Grok 4 family, consumer chat, or an API model ID. Those are not interchangeable for access or pricing.
Free chat is not the same as a free API
The plan page lists a Free plan at $0 per month with limited AI usage. It does not specify in the retrieved plan details whether the original Grok 4 model is available on that plan, or give a fixed message quota or reset schedule. (xAI plans)
At launch, xAI said Grok 4 was available to SuperGrok and Premium+ subscribers and through the API. A later announcement said Grok 4.1 was available to all users on Grok’s web, X, and mobile app experiences. That is an announcement about Grok 4.1, not proof that the original Grok 4 remains available at no cost today. (Grok 4 announcement; Grok 4.1 announcement)
| What you want to use | What the cited information establishes |
|---|---|
| Grok consumer chat | A $0 plan exists, but usage is limited. The cited plan details do not specify an exact allowance or confirm access to the original Grok 4 model. |
| Original Grok 4 in consumer chat | Its launch announcement listed paid subscriber plans. Current free access to the original model is unconfirmed. |
| Production API | API usage is billed separately from consumer plans. The original grok-4-0709 ID was retired and requests to it redirect to Grok 4.3; the cited retirement notice does not establish the applicable rate for redirected requests. (xAI plans; retirement notice) |
An August 10, 2025 post from @SpaceXSI on X claimed Grok 4 was free worldwide through Auto or Expert modes, with temporary “generous” limits. It gave no quota or reset interval, and the post alone does not establish that the account was authorized to announce xAI policy or that the offer remains in effect. Treat it as a historical claim, not current access guidance.
What production API use costs
xAI’s pricing table states that prices are in USD and lists text API rates per 1 million tokens. For grok-4.7, it labels the columns as input, cached input, and output for both short and long context. The short-context rates are $2.00, $0.50, and $6.00 respectively; the long-context rates are $4.00, $1.00, and $12.00. The listed long-context threshold is 200,000 tokens. When a prompt reaches the model’s long-context threshold, xAI says the long-context rates apply to all tokens in the request. (docs.x.ai)
| API model and pricing tier | Input / 1M tokens | Cached input / 1M tokens | Output / 1M tokens |
|---|---|---|---|
| Grok 4.7, short context | $2.00 | $0.50 | $6.00 |
| Grok 4.7, long context (prompt reaches 200K-token threshold) | $4.00 | $1.00 | $12.00 |
Illustrative short-context calculation: If a request stays below the long-context threshold and uses 100,000 uncached input tokens and generates 20,000 output tokens, the listed token charges total $0.32: (100,000 ÷ 1,000,000 × $2.00) + (20,000 ÷ 1,000,000 × $6.00). This assumes no cached input or separate tool charges. Actual charges depend on token usage, caching, whether the prompt reaches the long-context threshold, and any tool charges. (docs.x.ai)
If your code still sends requests to the retired grok-4-0709 slug, xAI says those requests are redirected to Grok 4.3. The available retirement excerpt confirms the redirect but does not establish the rate applied to redirected requests. Do not assume either that the old slug still serves the original model or that its replacement is billed at a particular rate without checking the pricing documentation. (xAI retirement notice)
The API quickstart instructs developers to load credits to start making API requests, and the plan page says API usage is billed separately rather than drawn from a subscription’s usage pool. The cited information does not confirm a free production API allowance. (API quickstart; xAI plans)
Check API capabilities against your workload
For Grok 4.7, xAI’s developer guide lists the API model identifier grok-4.7, a 500,000-token context window, text and image input, text output, and tools including function calling, web search, X search, and code execution. These are Grok 4.7 API details; do not assume they describe the retired original Grok 4 model or every consumer chat plan. (Grok 4.7 developer guide)
Before choosing a model, match its documentation to your application’s actual needs: image input, a particular tool, long prompts, or another required capability. Estimate spend using expected uncached input, cached input, and output tokens, and verify which context tier applies. If you specifically need the original Grok 4 API model, the retirement notice says its old API slug no longer serves that model. (Grok 4.7 developer guide; retirement notice)
Bottom line
- For no-cost experimentation: xAI lists a $0/month consumer plan with limited AI usage, but the cited plan details do not confirm access to original Grok 4 or state an exact quota. (xAI plans)
- For production API use: Treat API usage as separate from consumer plan usage. The retired
grok-4-0709ID redirects to Grok 4.3, but confirm the billing treatment for redirected requests before budgeting. (xAI plans; retirement notice) - For a reliable estimate: Use the exact model ID, expected token volumes, and confirmed context-tier rules. Do not use a third-party claim about Grok 4 Fast or Grok 4.1 Fast as pricing evidence for a different model.