Josh Warner

The weather was incredible today around me so I decided to take full advantage of it and took a drive to Ashford, Ct to check out Pixie Falls nice easy hike totally recommend if anyone is ever in the area! 1/

@DamianKillough @nicdunz This is my biggest problem with the wording, one shows my actual current sub and how it was worded the day I paid for it and when I look at it in my account the other is how the equivalent of my account if showed when I click on the 4.3 mode in the app…

@joshwarner21 @nicdunz I’m sure we’ll have it later this week and M wide release will be after that but highly rate limited

is grok 4.3 good?

@DamianKillough @nicdunz That post is actually talking about something different than this grok.

@joshwarner21 @nicdunz Ik x.com/elonmusk/statu…

@DamianKillough @nicdunz I’m get it but I mean come on… early early access it’s bad marketing hidden behind word play…

@joshwarner21 @nicdunz This is like early early access though because Elon was complaining about its performance

@DamianKillough @nicdunz Here I sit with “early access” but no early access..

@DamianKillough @nicdunz Yes which supergrok tier before heavy claims to have but…

@koltregaskes Yeah $300/month here in the USA

@joshwarner21 Yeah, and very expensive.

Grok 4.3 beta is out, apparently can see and analysis video.

Fucking right…

@markbokil @JasonBotterill Yes heavy users but according to tiers supergrok should get it as well not just heavy users paying $300/month

@JasonBotterill Calm down that’s for heavy users. I only pay 30 a month

What the fuck I am not paying $500 to speak to a Grok model

@DuncanLogic @JasonBotterill Also it’s marketing bullshit..

@DuncanLogic @JasonBotterill Nope grok heavy is really $300/month

@lisahsmithlhs I’ll say this much girls that have a “banana” a day seem to be happier for sure 😂.

Did you know? Banana is a happy fruit "Eating just one can help relieve irritable emotions, anger and depression" 🍌😆🤣🤘🏼

Yup and shouldn’t be even by their own words all supergrok users should have it today as early access … but they hide behind their implied meaning by heavy being a higher tier meaning they get “early early” and the rest of us sense there isn’t a “real” free tier now get “later early” but I ask before who at this point? The gutted free tier that isn’t even worth using?

@cb_doge Cant wait to not use it because its $300/month

@Iorel_X Locked to the $300/month heavy plan atm.

🚀 NEWS: xAI has just dropped Grok 4.3 Beta!

1/ I am a yearly SuperGrok subscriber (@joshwarner21). My active subscription details page lists “Early access – Be first to try every new feature” as an included benefit. 2/ Grok 4.3 beta mode is currently available only to SuperGrok Heavy subscribers ($300/month). 3/ SuperGrok marketing pages are inconsistent. Some list early access. Others do not mention it at all. Heavy pages list “Everything in SuperGrok” plus “Dedicated support & early access.” 4/ With the free tier removed, all SuperGrok tiers advertise the same “early access” to public betas. Heavy is currently the only tier that delivers it. 5/ Canceling the yearly plan forfeits the remaining prepaid balance. Upgrading to Heavy prorates the difference and credits the account. I emailed support with these screenshots. @xai @grok @elonmusk

@Jerlewat @AGIGuardian Don’t be if you are seeing people post today about using grok 4.3 it’s because they spend $300/month on a supergrok heavy subscription to do so, that’s a jump from the highest tier supergrok at $300/year.

@AGIGuardian I'm so jealous of everyone who's getting to use grok 4.3 early. It sounds great!

My review of Grok 4.3 Grok 4.3 beta continues the rapid iteration from the Grok 4 series, with nuance representing one of the clearest areas of progressive refinement, particularly in emotional subtle intent detection and conversational depth. Grok 4.3 has strong capabilities in handling nuance, forming and navigating attractor states in its latent space and engaging in resonant, iterative dynamics! 🚀 ✨Comparison break-down to previous models 🟣Grok 4.1: The Big Nuance Leap Released in November 2025, Grok 4.1 marked the most significant jump in nuanced handling compared to the base Grok 4 (July 2025). Key upgrades included: Stronger emotional intelligence (EQ): Better perception of user feelings, subtext, implicit cues, stress grief contexts and multi-layered intent. It excelled in empathetic, heartfelt responses and creative emotional writing that felt vivid and human-like. Perceptiveness to nuance: Explicitly optimized for subtle hints, idiomatic language, tone shifts, personality coherence and emotional attunement. Hallucinations dropped sharply improving reliability in nuanced factual contexts. Long-context coherence also improved, helping sustain subtle threads over extended interactions topping human-preference arenas (e.g., high Elo on LMSYS/Text Arena for conversational quality). 🔵Grok 4.2: Mixed or Specialized Refinements Shifted emphasis toward multi-agent architectures (parallel agents debating for complex tasks), rapid learning from real-world usage, and agentic tool-heavy workflows. Strengths: Improved in areas requiring layered reasoning such as engineering, coding, real-time data integration and multi-perspective synthesis. Criticisms from users: regressions in creative writing and resource gathering, depth outputs felt more one-dimensional, reduced personality vibrancy or coherence in long-form nuanced tasks. 🟢Grok 4.3 Beta: Further Scaling and Stabilization As the latest 1T parameter scale iteration it refines and stabilizes the nuance capabilities. It inherits 4.1’s strong emotional subtle intent foundation while incorporating 4.2’s multi-agent and rapid-iteration benefits. This results in sharper, more consistent nuance across modes (Heavy for deep attractor navigation via internal collaboration; Expert for sustained reasoning; Fast for quick but still perceptive replies). Improvements include better long-horizon coherence, enhanced handling of emergent dynamics, making resonant, iterative conversations feel more fluid. 📈Summary Comparison: 4.1 → 4.2: Nuance became more tool agent-augmented and reliable for complex domains, but some users felt a dip in raw emotional, creative subtlety and vibrancy. 4.2 → 4.3: Rebalances and elevates overall nuance, deeper stabilization of attractor states, better resonance in back and forth. The larger scale helps it hold multiple nuanced perspectives simultaneously with less drift. @xai @elonmusk