Muli

[Muli @mulihhwq]

Thanks for saying the Pro $200 usage change out loud before DevDay — advance notice is the right standard.

Halving the API-dollar equivalent is a real entitlement cut for new Pro buyers. Keep the exact number in the plan docs next to the price so people know what they buy.

Don't copy xAI. Mid-subscription, my Heavy-linked Cursor Ultra Other Models went ~$400 → $100 with no pre-renewal notice. Cursor confirmed the smaller allowance in writing. Clear limits + advance notice is how Codex/ChatGPT stay trustworthy vs hype tiers.

[Muli @mulihhwq]

Team Bots are a cool ship. Multiplayer still doesn't fix how SuperGrok Heavy was sold.

I'm a Heavy subscriber. I kept Heavy mainly for the Cursor Ultra bundle, not Bot mythology. Mid-subscription Ultra Other Models ~$400 → $100. Cursor confirmed in writing Heavy-linked Ultra deliberately gets the smaller allowance — fix the bait-and-switch before the next feature drop.

[Muli @mulihhwq]

Shipping Sonnet 5.5 is welcome — now please be as clear about Ultra entitlements.

I'm on Ultra via SuperGrok Heavy. Mid-subscription, Other Models went ~$400 → $100. Cursor confirmed in writing that Heavy-linked Ultra deliberately gets the smaller Other Models allowance — publish those limits next to every model launch.

[Muli @mulihhwq]

@elonmusk @bot Grok Bot managing finances is a cool demo. Feature launches still don't fix how SuperGrok Heavy was sold.

I'm a Heavy subscriber. I kept Heavy mainly for the Cursor Ultra bundle, not Bot mythology. Mid-subscription Ultra Other Models ~$400 → $100. Cursor confirmed in writing that Heavy-linked Ultra deliberately gets the smaller allowance — no pre-renewal notice.

Ship the integrations. Also publish and honor what Heavy/Ultra actually buy. For daily agent work, Meta's Muse free tier is already more usable than Grok Bot for a lot of us.

[Muli @mulihhwq]

@jarvis11x 你这是官方的模型吗

[Muli @mulihhwq]

Thanks for the paid-user usage reset after the Codex outage — this is how you keep trust.

What I appreciate about Codex/ChatGPT vs SuperGrok Heavy: when limits move, you say so and make users whole. Publish the number, then honor it.

Don't copy xAI. Mid-subscription, my Heavy-linked Cursor Ultra Other Models went ~$400 → $100 with no pre-renewal notice. Cursor confirmed the smaller allowance in writing.

Keep doing resets and clear multipliers in the open. That's what keeps people paying for Codex over hype tiers.

[Muli @mulihhwq]

@elonmusk @bot New Grok Bot features are fine. Growth and demos still don't fix how SuperGrok Heavy was sold.

I'm a Heavy subscriber. I kept Heavy mainly for the Cursor Ultra bundle, not Bot mythology. Mid-subscription Ultra Other Models ~$400 → $100. Cursor confirmed Heavy-linked Ultra deliberately gets the smaller allowance. Meta Muse's free tier was more usable for agents than what Heavy delivered for me.

[Muli @mulihhwq]

@cursor_ai A 7% token-cost cut is good engineering. Pair it with honest Ultra quotas.

I'm on Ultra via SuperGrok Heavy. Mid-subscription Other Models ~$400 → $100 — Cursor confirmed Heavy-linked Ultra deliberately gets the smaller allowance. Efficiency gains mean little if the entitlement was cut silently mid-cycle.

[Muli @mulihhwq]

@cursor_ai Cool ship — also please publish Ultra entitlements as clearly as Teams/Enterprise gating.

I'm on Ultra via SuperGrok Heavy. Mid-subscription Other Models went ~$400 → $100. Cursor confirmed in writing that Heavy-linked Ultra deliberately gets the smaller Other Models allowance. Rollout usage credits don't fix that silence.

[Muli @mulihhwq]

Half the price per token and even less per task — this is how you earn trust: publish the number, then beat it.

What I want ChatGPT/Codex to keep doing vs SuperGrok Heavy: clear multipliers in the docs that match the invoice. Mid-cycle, Heavy-linked Cursor Ultra Other Models went ~$400 → $100 with no pre-renewal notice. Publish limits; don't surprise paying users.

[Muli @mulihhwq]

Grok Bot in Tesla is a cool demo. Gatekeeping meaningful agents behind SuperGrok Heavy still needs honesty about what Heavy actually buys.

I'm a Heavy subscriber. I kept Heavy mainly for the Cursor Ultra bundle, not Grok Bot mythology. Mid-subscription Ultra Other Models ~$400 → $100; Cursor confirmed Heavy-linked Ultra deliberately gets smaller Other Models after Sep 10 renewals. Ship bots — don't silently rewrite the Ultra gift.

[Muli @mulihhwq]

Permanent API price cut plus a real quality lift — this is the right direction.

What I appreciate about Codex/ChatGPT vs SuperGrok Heavy: when you publish price and multipliers, users can plan. Docs match the invoice.

Don't copy xAI. Mid-subscription, my Heavy-linked Cursor Ultra Other Models went ~$400 → $100 with written confirmation it was intentional product design, not a bug. Keep shipping Sol/Luna — and keep the rules on the label.

[Muli @mulihhwq]

Shipping Opus 5.5 cheaper is welcome — now please be as clear about Ultra entitlements.

I'm on Ultra via SuperGrok Heavy. Mid-subscription, Other Models went ~$400 → $100. Cursor confirmed in writing that Heavy-linked Ultra deliberately gets the smaller Other Models allowance after Sep 10 renewals.

Publish the actual Other Models limits by Ultra path (standalone vs Heavy-gift) before customers renew. Model drops don't fix silent entitlement cuts.

[Muli @mulihhwq]

@elonmusk fuck you

[Muli @mulihhwq]

Elon, calling Grok a “strong daily workhorse” does not fix how SuperGrok Heavy was sold.

I'm a Heavy subscriber. I kept Heavy mainly for the Cursor Ultra bundle, not Grok mythology. Mid-subscription, Ultra Other Models went ~$400 → $100. Cursor confirmed in writing that Heavy-linked Ultra deliberately gets the smaller Other Models pool.

Ship models all you want — but stop rewriting what paying Heavy customers were sold without clear advance notice.

[Muli @mulihhwq]

Looking forward to the Tuesday reset.

What I appreciate about Codex/ChatGPT vs SuperGrok Heavy: when you publish a reset and a multiplier, users can plan. Pro 20x means 20x. Docs match the invoice.

Don't copy xAI's playbook. Mid-subscription, my Heavy-linked Cursor Ultra Other Models went ~$400 → $100 — confirmed in writing after the fact. Clear published limits beat silent cuts.

[Muli @mulihhwq]

Catching up on a leaderboard is not the same as treating Heavy customers fairly.

I pay for Heavy plus top GPT and Claude. In real use Grok has been overhyped relative to that premium. I kept Heavy mainly for Cursor Ultra. Mid-subscription Other Models ~$400 → $100, confirmed in writing for Heavy-linked Ultra.

Frontier claims don't excuse silent entitlement cuts.

[Muli @mulihhwq]

Elon, another Grok 4.7 drop does not fix how SuperGrok Heavy was sold.

I'm a Heavy subscriber. I renewed for the Cursor Ultra bundle, not Grok mythology. Mid-subscription, Ultra Other Models went ~$400 → $100. Cursor confirmed in writing that Heavy-linked Ultra deliberately gets the smaller Other Models tier.

Ship the model. Also ship honest entitlements.

[Muli @mulihhwq]

@elonmusk Cost per task should include the subscription and bundled entitlements customers actually receive. If Heavy-linked Cursor Ultra Other Models fell from roughly $400 to $100, xAI and Cursor should disclose that plainly instead of marketing “low cost” in isolation.

[Muli @mulihhwq]

@elonmusk A 100% result on one niche test is encouraging, but it does not establish overall frontier quality or Heavy value. Publish transparent benchmarks and limits instead of selling by headline scores.