From community reports, SuperGrok Heavy seems to offer significantly less than top-tier competing plans like Claude Max 20x and ChatGPT Pro 20x, especially when looking purely at token usage and credit value as a multiplier of the subscription price. But I don't think that's the main issue. OpenAI and Anthropic keep improving. Their models are getting smarter and cheaper. Meanwhile, SpaceXAI is losing its strong position on performance and cost-effectiveness. Grok 4.5 used to have very good cost per actual work done compared to OpenAI and Anthropic models at the time. I think that's a big reason many users moved to Grok. The quality was good and it felt worth the price. Grok 4.7 doesn't feel the same. This also lines up with Artificial Analysis benchmarks and community feedback that both SuperGrok Heavy and Cursor Ultra quotas drain much faster since Grok 4.7 launched. I think the SpaceXAI and Cursor teams are amazing, and I believe Grok 4.8 or Grok 5 will fix a lot of this and get smarter again. But I also hope it becomes as cost-effective as Grok 4.5 was, or that SpaceXAI introduces another model focused more on cost-effectiveness. Right now, SpaceXAI is missing a small, cheap, high-performance model comparable to GPT-6 Luna. Both Grok Build and Cursor end up using top-tier models even for simple tasks, burning more quota than necessary. Yes, I know we can connect custom PAYG models to both Grok Build and Cursor and use cheaper models for those tasks. But for a $300 subscription that's already 50% more expensive than competing $200 plans, I don't think it makes sense that I still need to pay seperately for another model just to handle simpler tasks. $300 is not a starter tier. It's a high-end subscription right now, which requires a very high willingness to pay. Many users are complaining not because they hate Grok, but because they like it, had high expectations for it, and feel disappointed with how things are right now. Hopefully someone on the teams is listening to community feedback.