Grok, Grok Bot, Imagine and the API: products and costs compared

Understanding xAI, SpaceXAI and SpaceXSI

SpaceX announced its acquisition of xAI on February 2, 2026. Current branding uses SpaceXSI; the privacy policy identifies SpaceXAI LLC. We distinguish product names, branding and contracting entity. An older reference to xAI or SpaceXAI does not automatically describe a different Grok service.

For your decision, record the product used, account and applicable terms. A general reference to Grok is insufficient for technical or privacy acceptance. Our content describes independent implementation by lol.marketing LLC and does not imply a manufacturer partnership.

Grok chat: work through a task together

For an initial trial, choose a bounded task: summarise approved material, organise arguments or prepare questions about a document. Supply the desired output and authoritative sources. Then check whether the answer reflects those sources accurately and leaves uncertainty visible.

Grok on X has its own usage and privacy context. Do not assume settings from another product apply to it. For work, agree which account may be used and which information may be entered. A convenient starting point does not by itself establish permission to process customer data.

Grok Bot: an AI agent with handoffs and reviewed results

According to the product overview, Grok Bot is an AI agent that works on a persistent cloud computer with a browser, file system and terminal; the apps run on macOS, Windows, Linux, iOS and Android (checked October 25, 2026). A Bot project should define the work to be carried forward and when a person intervenes. For an internal briefing, that might mean using approved files, collecting changes with references and saving a draft for review. We evaluate the result, including unresolved issues.

For ongoing roles and recurring workflows, our implementation service explains the route from account review to training. You decide the commissioned scope. A process without clear sources and acceptance does not become dependable simply because a Bot takes it over.

Imagine: media production as a distinct workflow

The Imagine documentation covers image and video generation and editing. Selection needs more than that feature list: an abstract editorial motif requires different checks from a product depiction. Define which details must be accurate, which references are permitted and who approves publication.

Our media workflow guide covers briefs, variations, rights review and filing. Do not treat generated images as evidence of real events or customer outcomes. Choosing an appropriate visual remains an editorial and domain decision.

Build, models and the API: plan development deliberately

The Build announcement describes a coding agent. The current API model card lists Grok 4.7 with 500,000 tokens of context and text/image inputs. These are technical capabilities rather than a finished business application. Integration also requires validation, access control, error handling and operation.

Grok Voice 3.0 was introduced on October 14, 2026. A voice project should define conversational purpose, human handover and misunderstanding handling. The announcement does not establish suitability for your telephony system, language or industry. Those require a separate scoped test.

What does Grok cost? Plans, usage and external costs

As checked on October 25, 2026, the provider pricing page lists Free at USD 0 and Plus at USD 8 per month, neither including Grok Bot. Grok Bot is included from Premium at USD 30; Super costs USD 100 and Ultra USD 200 per month. Taxes and fees may apply. The subscription documentation explains shared weekly usage pools and separate legacy arrangements.

For your pilot, compare subscription, expected usage, integration, review effort and support. A low individual-call price says little about a complete process requiring several attempts. Include correction and waiting time. Quantified savings should come from your own comparison data. Our service fees are agreed independently of the manufacturer offer.

Four questions for the next step

Evaluate a pilot with a short decision table: task, permitted data, achievable output, required review and remaining uncertainty. Apply the same criteria to every option considered. If an option depends on additional manual steps, include them in the comparison. The choice then remains understandable to colleagues who did not attend the demonstration.

Answer these questions for one workflow first. The linked workflow scorecard provides a structured way to prepare several ideas. Discuss concrete Bot requirements through our independent implementation service. This comparison records the source-check date; it does not replace reviewing your account. Implementation and support are provided independently by lol.marketing LLC. We are not affiliated with, authorised by or endorsed by xAI or SpaceXSI.

• Which specific input should produce which usable output?
• Who reviews content, data access and external effects?
• Which interface is actually available in the approved account?
• Which failure cases must pass before proceeding?