xAI Grok — Frontier Chat Model With Native X Firehose + Cheap API (xAI Folded Into SpaceX as “SpaceXAI” 2026, Grok Brand Retained, Active)

X.AI (founded Nov 2023 by Elon Musk, Palo Alto) was acquired by SpaceX Feb 2 2026, dissolved as independent entity May 7 2026, rebranded “SpaceXAI” Jul 2026 — per your Roam Around→Layla rule this is acquired-kept: Grok model family, grok.com, and api.x.ai all still live (old model slugs auto-redirect to Grok 4.3), brand “Grok” retained, not shutdown, not a silent rebrand of the SKU. Active 2026-08. Category ai-productivity-data; pipeline_stage none. Killer differentiator = native real-time X/Twitter firehose + DeepSearch sentiment reads no rival matches + Grok 4.6 at $2/$6 per M tokens (under half Claude Opus 5 API) + 500K context + Grok Imagine video/voice/Build agents; not ChatGPT (broadest ecosystem, no X firehose), not Claude (safer/prose/coding reliability, pricier), not Gemini (Google-stack native), not Perplexity (pure citation engine, no image/video/agent). 2026 trap: tier sprawl (Free/X Premium/SuperGrok Lite/SuperGrok/SuperGrok Plus/SuperGrok Heavy/X Premium+ = 7 consumer ladders), Jun 2026 switched to shared weekly usage pool (old per-2h caps gone), >200K token prompts bill 2× on API ($4/$12), imagine/video draws from same weekly pool, iOS no system push, guardrails loosest of frontier labs (brand-safety risk), SpaceXAI endpoint migration TBD 12+ mo.

1. One-line positioning

Grok = real-time-social-native frontier assistant: grok.com + iOS/Android apps + X-integrated pane → prompt with live web/X DeepSearch (40+ sources, <1 min reports) → Big Brain extended reasoning → Grok Imagine image/video (Aurora + Imagine, 480p/720p, audio) → Voice mode → Grok Bot AI teammates (Heavy-only beta, signs into tools, cloud-machine per bot, bot-manages-bot) → Grok Build coding CLI/TUI (Cursor-trained, GitHub Copilot dropped Grok Code Fast 1 May 15 2026) → xAI API (OpenAI-compatible, server-side web/X/code/file search, RAG built-in Dec 2025). Fits ai-productivity-data; pipeline_stage none (prompt→search→reason→generate→tool-call→publish). For X-centric marketers / cost-sensitive API builders / real-time sentiment watchers, it’s the only native seat; for regulated enterprise or deepest prose, Claude/ChatGPT win.

2. Core features

Layer	Capability	Notes
Models (Aug 2026)	Grok 4.6 flagship (500K ctx, $2/$6, Aug 12 2026), Grok 4.5 (500K, same price), Grok 4.3 (1M ctx, $1.25/$2.50, native video input), Grok 4.20 reasoning/non-reasoning/multi-agent, Grok 4.1 Fast ($0.20/$0.50, 2M ctx), Grok Build 0.1 (coding $1/$2)	Legacy grok-3/4/4.1-fast retired May 15 2026 → redirect to 4.3
DeepSearch	Live web + X firehose, 40+ sources, inline citations, <1 min synthesis, DeeperSearch preset	Free 10/day, paid 20/2h (now within weekly pool)
Big Brain	Extended reasoning mode, configurable effort none/low/med/high	API bills reasoning tokens at output rate
Grok Imagine	Aurora images + Imagine video (480p/720p, audio); consumer draws from weekly pool; API $0.02–0.07/img, $0.05–0.07/sec video	Loosest content filters of frontier labs
Voice	Spoken chat, TTS $15/M chars, STT $0.10–0.20/hr, Voice API $3/hr	No iOS system-level push
Grok Bot (beta)	AI teammates: cloud machine per bot, observes user, bot-manages-bot, signs into inboxes/CRMs; Heavy/Cursor Ultra only	Launched Aug 11 2026, no independent reli test yet
Grok Build	Coding agent CLI/TUI, Cursor-trained, code execution sandbox	Replaces Grok Code Fast 1 (deprecated in Copilot May 15 2026)
API surface	`api.x.ai` OpenAI-compatible, cached input ($0.30–0.50), server tools web/X search $5/1K calls, code exec $5, file $10, collections $2.50; Live Search $25/1K sources	Endpoint migrates to SpaceXAI URL within 12+ mo, no break yet
Enterprise	Grok Business $30/seat (RBAC, no-train opt), Enterprise custom (SSO/SCIM/CUME/audit) launched Jan 2026	Azure AI Foundry deploy since Sep 2025

3. Pricing (2026-08, USD)

Consumer (7 tiers, shared weekly usage pool since Jun 2026):

Tier	Monthly	Annual	Model access	Notes
Free	$0	$0	Grok 4 Mini (text+web+X), no Imagine	Separate chat/voice caps, own reset
X Premium	$8	—	Grok 4 inside X, tight caps	Buy only if you want X blue/ads
SuperGrok Lite	$10	—	Grok 4, 1 agent, 480p/6s video	Mar 25 2026 entry tier
SuperGrok	$30	$300 ($25/mo)	Grok 4.5/4.6 rolling, DeepSearch, Big Brain, Imagine, Voice	Main paid seat
X Premium+	$40	—	Same + X ad-free/perks	Only if X-native
SuperGrok Plus	$100	—	Higher weekly pool, 1080p video, priority peak	For $30-ceiling users
SuperGrok Heavy	$300	$3,600	Full Grok 4.5/4.6 + Grok 4 Heavy multi-agent + Grok Bot	Researcher/heavy agentic

API (pay-per-token, no plan needed):

Grok 4.6 / 4.5: $2.00 in / $6.00 out per 1M, 500K ctx, cached $0.30 (4.5) / $0.50 (4.6); >200K prompt → $4/$12 whole request
Grok 4.3: $1.25 / $2.50, 1M ctx, >200K → doubles
Grok 4.1 Fast: $0.20 / $0.50, 2M ctx
Grok Build 0.1: $1 / $2
Batch API 20% off; server tools extra (web/X $5/1K, code $5, file $10)

pricing_type = freemium-plus-x-bundle-plus-per-token-api; starting_price = $0; free_tier = "yes; Grok 4 Mini, real-time web+X search, limited voice, NO Imagine, separate reset cycle, no paid pool"; pricing_note = "feb-2026-spacex-acquired-xai-may-2026-dissolved-rebrand-spacexai-grok-brand-kept-acquired-kept; jun-2026-paid-tiers-shared-weekly-usage-pool-not-per-2h-caps; supergrok-30-25-annual; heavy-300-only-full-4.6-4-heavy-grok-bot; api-grok-46-2-6-500k-context-over-200k-bills-4-12; grok-3-4-4.1-fast-retired-2026-05-15-redirect-4.3; imagine-video-draws-weekly-pool-consumer; endpoint-migrates-spacexai-url-12mo-window-no-break-yet"

4. Access Type

access_type: web-app-plus-mobile-plus-api

access_display: grok.com · 📱 iOS/Android · 🔌 X.com pane · api.x.ai (OpenAI-compatible) · Azure AI Foundry · Cursor/Grok Build CLI · no MCP server documented, no native HubSpot/Salesforce connectors

5. Reviews (2026)

Source	Read
Stackaible / AI Rankings	“Situational yes — X firehose native, 4.6 frontier-class at half Claude Opus 5 API cost; ecosystem/integrations trail OpenAI/Anthropic”
G2/Capterra	Not G2-listed as traditional SaaS; app-store ratings ~4.3 iOS, complaint axis = tier confusion + Imagine throttling after 50–100 gens
Trustpilot	Mixed; billing cancels less angry than Semrush/Ocoya, but “SpaceXAI rebrand broke my bookmarked docs” confusion
Independent bench	Grok 4.6 AA Intelligence Index 61 (level with GPT-5.6 Sol), under Claude Opus 5 on coding reliability, over all on X-sentiment

Praise: unmatched live X signal, DeepSearch <1 min vs ChatGPT 30 min, $2/$6 API undercuts frontier, Grok 4.3 1M ctx + native video input, Imagine loose filters = good for rapid mockups, Big Brain solid on math.

Gripes: 7 consumer tiers confuse buyers, weekly pool opaque (no per-feature meter), >200K token 2× API bill surprises, Imagine soft-cap throttles, guardrails loosest (brand/safe-for-work risk), Grok Bot Heavy-only + unproven, SpaceXAI migration clouds long-term endpoint ownership, no MCP/native CRM, non-X research citation trails weaker than Perplexness.

6. Best for / Not for

Best for	Not for
X-native marketers / creators tracking real-time sentiment & trends	Regulated industries (finance/health/legal) needing strict audit + guardrails
Cost-sensitive API builders (high-volume $2/$6 or $0.20/$0.50 Fast)	Teams wanting deepest prose/long-form (Claude)
Researchers doing live-event synthesis (DeepSearch + X)	Non-X shops where firehose differentiator evaporates
Coding agents on Grok Build / Cursor (price-per-token)	Enterprises needing SSO/SCIM today without waiting SpaceXAI transition
Heavy agentic multi-model fans (SuperGrok Heavy $300)	Casual users who won’t use X/Imagine/Voice — Free is enough

7. Competitors

Tool	Lane	Starts
ChatGPT (OpenAI)	Generalist default, GPT-5.6, broadest ecosystem	$20/mo, $5/$30 API
Claude (Anthropic)	Prose/reasoning/coding reliability, safest	$20/mo, $5/$25 API
Gemini (Google)	Google-stack native, 1M+ ctx, Veo image/video	$20/mo, $1.25–2.5 API
Perplexity	Citation-first answer engine, no image/video	$20/mo
DeepSeek	Cheapest API ($0.14/$0.28), no X firehose	$0.14/$0.28
Mistral/Cohere	Enterprise EU/US alt	custom