What actually changed (before vs after)
The Fall 2026 SuperGrok repricing changed three things: Lite price, Plus tier introduction, and per-tier volume ceilings.
Before (early-to-mid 2026):
| Tier | Cost/mo | Images/mo (attempted) | Videos/mo (attempted) |
|---|---|---|---|
| SuperGrok Lite | ~$10 | ~600 | ~90 |
| SuperGrok (middle) | $30 | ~1,500 | ~240 |
| SuperGrok Heavy | $300 | ~10,000 (est.) | ~2,000 |
After (Fall 2026):
| Tier | Cost/mo | Images/mo (attempted) | Videos/mo (attempted) |
|---|---|---|---|
| SuperGrok Lite | $30 | ~1,500 | ~240 |
| SuperGrok Plus (new) | $100 | ~4,500 | ~720 |
| SuperGrok Heavy | $300 | ~10,000 (est.) | ~2,000 |
The old middle-tier SuperGrok at $30 was effectively renamed to "Lite" and repositioned as the entry. The new Plus tier at $100/mo fills the gap between the entry level and the pro Heavy tier. Heavy stayed unchanged.
X Premium ($8/mo) and X Premium+ ($40/mo) still exist as bundled X social subscriptions with some Grok access. Those didn't change with the SuperGrok repricing — they're separate SKUs selling X social features primarily.
Why xAI restructured (best guess)
xAI didn't publish detailed reasoning for the repricing. The best community-consensus read is a combination of three pressures:
Infrastructure cost accounting. As xAI became a SpaceX subsidiary in early 2026, pricing decisions increasingly look like they're being made against unit-economics targets rather than growth-at-all-costs. Image and video generation compute costs are meaningful per render — tripling Lite from $10 to $30 while giving 3x the volume means the per-user compute margin stays roughly constant while the price sensitivity ceiling moves up.
Tier ladder engineering. The old two-tier ladder (Lite at $10, Heavy at $300) had a $290 gap between them. Users hitting Lite's ceiling had no incremental option — jump 30x in price or downgrade. The new middle tier at $100 gives users a ~3.3x step instead of a 30x step. Standard SaaS pricing pattern.
Enterprise positioning for Heavy. With Plus at $100 filling the mid-tier, Heavy at $300 reads more clearly as an enterprise/prosumer tier rather than "the tier you upgrade to when Lite doesn't work." The pricing anchor shifts.
None of these are user-friendly rationales, but all three are structural. The pricing isn't going back down.
Lite went from $10 to $30 — better or worse value?
The Lite tier tripled in price but got 2.5x the image volume and 2.7x the video volume. On a per-attempted-render basis, Lite is roughly the same value ratio (slightly worse: 3x price for 2.5-2.7x volume). But that's not the useful comparison.
Better value if: You were an old Lite user generating close to the ceiling (~600 images/mo). The new Lite gives you 2.5x more volume at only 3x the price — you're paying $20 more monthly to unlock capacity you'd otherwise have to jump to the middle tier ($30 old) or Heavy ($300) for.
Worse value if: You were an old Lite user generating well below the ceiling (say ~100 images/mo). You're now paying 3x the price for capacity you weren't using anyway. The old Lite at $10 was the "occasional use" tier; new Lite at $30 is the "regular use" tier. Occasional users got pushed into paying for capacity they don't need.
Worse value across all cases: The moderation reject rate (~60% on romantic/companion prompts) and the throttle stack (24h + 2h burst + peak-hour) didn't change. Every new $30 is being spent against the same underlying friction.
The new Plus tier at $100 — who's it for?
SuperGrok Plus at $100/mo is the tier introduced in Fall 2026. 3x Lite's volumes at 3.3x the price. On per-render economics it's slightly worse value than Lite but with a lighter peak-hour throttle.
Best fit: Users who consistently hit Lite's ceiling (~1,500 attempted per month = ~50/day) AND generate during peak hours where the throttle degradation matters. Plus gets a lighter throttle than Lite, so if your usage clusters US 6-11pm ET weekdays, the throttle relief might justify the $70/mo step up.
Bad fit: Anyone whose usage sits somewhere in the middle (150-600 attempted per month). At that volume, Lite's ceiling isn't your bottleneck, and Plus's higher ceiling doesn't buy you anything you'll use. You're paying $70 extra monthly for headroom you don't reach.
Honest read on the new tier: it's a good option for a specific narrow slice of high-volume paid users. For most Lite users, it's not worth the upgrade. For Heavy users, the step down to Plus doesn't save meaningful money relative to the ceiling loss.
Heavy stayed at $300 — same product, clarified ceiling
SuperGrok Heavy stayed at $300/mo — no price change with the Fall 2026 repricing. What did change is the volume ceiling clarified: ~10,000 attempted images and ~2,000 attempted videos per month, plus Grok 4.5 access and highest-priority queue placement.
Heavy is priced against enterprise creative tools (Runway Enterprise, Midjourney Mega, Adobe Firefly for teams) and stayed in that range. Priority queue placement gives Heavy users the lightest peak-hour throttle degradation among all SuperGrok tiers — but "lightest" is not "none." Heavy users still hit the throttle during US evening hours; they just wait less than Lite/Plus users.
Heavy as a value proposition works if you're generating professional-scale volume (100+ images/day or 25+ videos/day). It doesn't work if you're generating occasional-to-moderate volume — you're paying for a ceiling you won't come close to hitting. Most personal users would be better served by Lite at $30/mo or Sloane Premium at $19.99/mo.
What DIDN'T change: moderation, throttling, identity drift
The repricing changed the price ladder. It didn't change any of the underlying friction points that make Grok Imagine feel worse than the sticker suggests.
Moderation reject rate. Grok's classifier still rejects ~60% of romantic/companion-shaped prompts, and rejected renders still count against your quota. This is the biggest hidden cost across every SuperGrok tier — you're paying for 1,500 images but usable output is ~600.
Throttle stack. Three layers: rolling 24-hour daily cap, 2-hour burst cap (fires at 2-3 gens per 10-minute window), undocumented peak-hour throttle US 6-11pm ET weekdays. Applies to every tier including the new $100 Plus.
Persona identity drift. Grok Imagine is stateless. Same prompt renders subtly different characters every time. No tier upgrade fixes this because the architecture doesn't have per-character anchoring.
If any of these were your specific pain point pre-repricing, the repricing didn't solve it. A higher price tier only helps if the daily ceiling was your pain point.
When the new SuperGrok tier structure is better value
New Lite at $30/mo is better value if: You were consistently hitting the old Lite's ~600-image ceiling and doing the "grind through pain" thing. New Lite's 1,500 attempted (~600 usable at 40% success) gives you the same usable output at $30 that you were previously getting via the old $30 middle tier. Effectively no repricing pain if you were already on the middle tier.
New Plus at $100/mo is better value if: You were on the old Heavy at $300/mo but not using the full ceiling. Plus gives you 3x Lite volumes at less than half of Heavy's price. If your usage sits at 1,800-4,000 attempted images/month with priority-queue benefit, Plus is a meaningful downgrade in cost.
Heavy at $300/mo (unchanged) is better value if: You're a professional generator running at 4,000+ usable images/mo where the ceiling is a real constraint and priority queue during peak hours matters for time-sensitive work.
For everyone else — casual users, occasional generators, companion-imagery users — the repricing made things worse. The floor moved up from $10 to $30. That $20/mo delta is meaningful for the exact use case (occasional casual generation) that Grok Imagine was best positioned for pre-repricing.
When Sloane's $19.99 flat wins outright
The Fall 2026 repricing pushed Grok's floor from $10 to $30. Sloane Premium at $19.99/mo now sits below the cheapest Grok tier while delivering a fundamentally different value proposition.
If you generate under 70 photos or 14 videos per month: Sloane Premium at $19.99 is cheaper than any Grok tier AND every render lands (100% success rate on intended-use content) AND same persona every gen AND never throttled AND chat + voice + memory + video all included on the same subscription.
If you generate 70-600 photos or 14-96 videos per month: Grok Lite wins on per-usable-render cost — IF you can tolerate the 40% success rate, three-layer throttle stack, and stateless identity drift. Sloane's answer is credit packs on top of Premium ($9.99/100, $19.99/250, $29.99/500). No commitment to $70+/mo you might not use.
If you generate 600+ photos or 96+ videos per month: Grok Plus or Heavy wins on raw cost per usable render. Sloane isn't built for this ceiling.
Sandra is a working example of what Sloane's flat pricing feels like: chat with her, ask her for a photo whenever, get an animation on Premium, no throttle stack, no moderation waste, no stateless identity drift on regens. If what you actually wanted from Grok Imagine was "consistent visual of the same girl I keep interacting with," Sloane solves it natively at a lower price point than any post-repricing SuperGrok tier.
