Grok review 2026: Grok 4.7, Imagine, Voice, Build and Bot, new pricing, pros, cons and who it suits

Grok has changed a lot in the past few weeks: a new flagship model, a new voice model and, on October 12, a new way to pay for it. This review covers what Grok does, how good SpaceXAI says it is, what it costs now and who should use it. It draws on SpaceXAI’s own announcements and pricing pages and on public reporting, and we haven’t run extended hands-on tests, so benchmark figures are the vendor’s.

Screenshot: the Grok web app at grok.com, signed out. Captured by Universal Digital Assistant on October 17, 2026.

Grok at a glance

What Grok can do

• Chat and voice. The Grok app opens on a prompt box with a model selector, and you can talk to it by voice. SpaceXAI calls Grok Voice 3.0 the world’s most intelligent speech-to-speech model, and says it handles millions of conversations a day in Tesla vehicles.
• Real-time answers. Grok draws on web search and on posts on X, which is its main difference from other assistants. It is also built into the X app.
• Imagine. Grok Imagine makes images and video; see What is Grok Imagine?. On the new plans, video generation starts at Premium.
• Build. Grok Build is SpaceXAI’s coding tool, available on the web and mobile, with project memory that carries conventions between sessions. Our best AI coding assistants guide covers it.
• Bot. Grok Bot runs persistent agents with their own computer that can start work without a prompt, and SpaceXAI has added Team Bots, shared AI teammates, and an Enterprise version.
• Cursor. Since October 12 the same plan includes Cursor; see our Cursor review.

Grok 4.7: what SpaceXAI claims

Grok 4.7 uses a new, larger base model with a longer reinforcement-learning run, and SpaceXAI says it works longer on difficult tasks and checks its own work more carefully. It is priced from $2 per million input tokens and $6 per million output tokens, the same as Grok 4.6, with a fast variant at twice the output speed for twice the price. SpaceXAI’s benchmark table compares it with Grok 4.6, GPT-5.6 Sol and Claude Fable 5.1:

The pattern is that Grok 4.7 is competitive with GPT-5.6 Sol and close to Fable 5.1 on some tests at a much lower token price, but behind Fable 5.1 on terminal work. These are vendor figures, and the comparison models are not the newest ones: OpenAI has since released GPT-6.1 Sol and Anthropic Claude Sonnet 5.5 and Claude Haiku 5.5. SpaceXAI also says Grok 4.7 has its strongest safeguard stack so far.

What Grok costs now

Grok is sold through The X Subscription:

Existing SuperGrok subscribers can keep their plan or switch with a prorated refund.

Pros

• Real-time answers with X data, which other assistants can’t match.
• Low token prices. $2 and $6 per million tokens is a fraction of rival flagship prices.
• A broad bundle. Chat, voice, images, video, coding agents and X features in one plan.
• Strong voice. SpaceXAI cites an independent index in which Grok Voice 3.0 ranks first.

Cons

• Vendor benchmarks only, and some rivals have newer models than the ones SpaceXAI compared against.
• Safety record. CNBC reported in January 2026 that xAI limited Grok’s ability to create sexualized images of real people on X after backlash, so check the current safeguards before relying on Imagine.
• Shared usage pool. Grok and Cursor draw from the same weekly allowance.
• Third-party models cost extra in Cursor.

Who should use Grok?

Grok suits people who want live information from X, creators who will use Imagine and X Premium features, and developers who want one bill for Cursor and Grok models. If you mainly want a general assistant with the widest app ecosystem, compare it with ChatGPT in our Grok vs ChatGPT page, and see the Grok profile.