THE PRODUCT
Grok by xAI
Dev YouTubers rate new Grok models highly, but user discourse is consumed by training-data and safety scandals.
THE VERDICT
REALITY SCORE · OUT OF 10 · CONFIDENCE HIGH
COMPOSED FROM
SENTIMENT · 244 REVIEWS
BEST PRICE TODAY
// Affiliate link — score is unaffected.
AT A GLANCE · QUOTABLE
- Rating: 6.5 / 10 (high confidence)
- User voices: 244 across 4 platforms
- Sentiment: 20% positive · 65% negative
- Updated: Sep 10, 2026
GYIBB rates the Grok by xAI 6.5/10 based on 244 user voices from 4 platforms. Confidence: high. Source: https://gyibb.com/ai-models/grok-by-xai
BUY IF
New Grok model rated genuinely capable by dev-audience YouTubers (Theo, 564k subs)
- + Aggressive datacenter/compute scaling seen by commenters as a credible path to top-tier performance
- + Grok Bot agents show real daily-use workflows for power users (How I AI, 114k subs)
- + Separate Chat/Build/Imagine subscriptions offer distinct specialized toolsets
SKIP IF
USER discourse dominated by Canadian Centre for Child Protection allegations of CSAM training data — severe trust damage
- − Safety/filtering failures acknowledged even by fans ('MechaHitler' self-designation, unfiltered image generation)
- − Fragmented subscription 'token buckets' across plans that run out weekly and monthly — confusing flat-rate UX
- − Owner-brand political backlash (Musk) spills directly onto the product in user communities
Where the layers disagree ⚡
6 CONTRADICTIONS DETECTEDUSER vs VIDEO: Lemmy users demand Grok's deletion over alleged CSAM training data, while VIDEO commenters (Theo, 564k subs) call the new model 'good' and predict it 'becomes the best quite quickly' — reputation and capability are being judged in completely different frames.
VIDEO partially aligns with USER safety fears: even pro-Grok comments on Theo's video note unfiltered output ('draw your entire classroom naked at once'), echoing the USER layer's 'MechaHitler' behavior outrage.
VIDEO vs flat-rate UX: How I AI's top comment describes fragmented token buckets across separate Grok Chat/Build/Imagine/Bot subscriptions that 'run out' weekly AND monthly — real subscription friction that the controversy-consumed USER sample never gets around to mentioning.
USER layer skew: zero of the top 25 user comments address pricing, latency, coding, or output quality, so there is no user-side corroboration — positive or negative — for the VIDEO layer's capability claims.
VIDEO internal split: Theo's 152k-view dev audience is bullish on capability, while alexusadays' 2.5k-view hands-on test drew an 'A whole night?' reaction to agent task speed — small-sample doubt about real-world agentic reliability.
No layer provides API economics: neither USER nor VIDEO data mentions $/1M-token pricing, rate limits, or API latency, so per-token value is entirely unevidenced in this dataset.
Value depends on how you pay ⚖
SAME MODEL · TWO BUYERSON A SUBSCRIPTION
6.5Claude Max · ChatGPT Plus · GLM Coding — flat rate, tokens don't bill
Flat-rate buyers get the strong half of the story: Theo's 564k-sub audience rates the new model highly and one '7 agents I use daily' video shows genuine daily utility. But the How I AI comment is a red flag — usable Grok Bot capacity reportedly requires stacking Chat, Build, AND Imagine subs, each with separate token buckets that run out weekly and monthly. Capability high; daily-limit architectu
ON PER-TOKEN API
4.5Enterprise / pay-per-use — $/1M, latency, token efficiency bite
Effectively unevidenced in this dataset: no comment — user or video — mentions $/1M-token pricing, API rate limits, or latency. The only adjacent signal is commenters crediting xAI's rapid datacenter buildout (favorable for throughput long-term, unproven now), and reputational/regulatory risk from the training-data controversy, which enterprises weigh heavily. Per-token buyers should seek benchmar
WHERE THEY AGREE +
WHERE THEY DON'T −
Where the 244 sources came from
VIEW EVERY CITATION →The four realities of the Grok by xAI
Most review sites collapse everything into one number. We keep the layers separate so you can see where reality bends.
What actual buyers say
What reviewers showed on camera
7 Grok Bot agents I use every day
How I AI · 271,659 views
"[comment] You should discuss Usage and tokens.....To have a capable Bot account with considerable weekly usage allowance, you must also have a VERY stacked Grok chat, Build, and Imagine subs. All with their own "token buckets" and usage me…"
Oh no (the new Grok model is good)
Theo - t3․gg · 152,355 views
"[comment] Google is like that meme of the turtle that only runs fast as hell when it gets scared. [comment] Theo is no longer hiding his 5.6 access [comment] Best model counter guy? [comment] Get the counter guy [comment] I have a feeling t…"
I Tried Grok Bot. Honest Review
alexusadays · 2,547 views
"[comment] thanks. very interesting. [comment] can it develop website by working on specific theme i install and can this use the theme option to setup the website for you through customizer? [comment] A whole night? [comment] don't do AI a…"
What the press said
What the brand says
no brand page found
* This page may contain affiliate links. No additional cost to you.
SIMILAR IN THIS CATEGORY
See all →
DATA SOURCES & AUDIT
244 data points across 4 platforms, synthesized via GYIBB's Truth Engine and fact-checked against source data before publication.
CONFIDENCE: HIGH · ANALYSED: SEPTEMBER 10, 2026 AT 10:25 AM · PROMPT V1.0 · READ METHODOLOGY →
Was this review helpful?
Embed this review
