Muse Code usage limit reached: what to do

What to do when you hit it

1. Wait for the refresh. The 5-hour allowance comes back every five hours. If the weekly limit is what ran out, you wait for the weekly reset; Meta does not publish the day or time.
2. Upgrade in the CLI. Type /upgrade in a Muse Code session to open Accounts Center, or go to Subscriptions > Muse Code there. An upgrade takes effect immediately and you pay a prorated amount for the rest of the cycle. High Usage ($15) is 5x Everyday, Power ($50) is 20x, Maximum ($200) is 80x.
3. Switch to pay-as-you-go for now. Any extra API key you create under your Meta Model API account is billed per token, not against your plan. If META_API_KEY is set, Muse Code uses it ahead of your browser sign-in, so unset it when you want your plan back. Muse Wonder costs $4 in, $0.20 cached and $20 out per 1M tokens; Muse Spark 1.3 is $1.25 in and $4.25 out. Muse Wonder Full is subscription only.
4. Make each prompt cheaper. /compact replaces earlier history with a summary, and /new starts a fresh session with token counts reset, so long threads stop resending everything. For routine edits, Muse Spark is the lighter pick: on the API it costs $1.25 per 1M input tokens vs $4 for Muse Wonder. Meta does not publish how each model counts against plan limits.
5. Check Contributor mode. Subscriptions start in Contributor mode, which gives more usage in exchange for letting Meta train on your prompts and code. If you turned it off, your plan now stretches less far.

Recent Muse Code resets

Meta has reset Muse Code limits twice since Muse Wonder launched on Oct 12. Any new one shows up on the live resets page within the hour.

Common confusions

• "There is no upgrade button": plan changes live in Accounts Center, not in a Muse Code settings screen. /upgrade opens it from the CLI.
• "I downgraded and still have the bigger limit": a downgrade starts at your next billing cycle, and you keep the current limit until then.
• "Do cached tokens count?" On the API, cached Muse Wonder input costs $0.20 per 1M vs $4 uncached. Meta does not say how cached tokens count against plan limits.
• "I pay for Meta Muse, why am I limited in Muse Code?" They are separate subscriptions. Muse Power and Maximum are bought at muse.ai or in the app; Muse Code plans are bought in Accounts Center with a Meta Model API account and only work through the Muse Code CLI.
• "Can I buy extra usage?" Meta lists no top-ups for Muse Code plans. Past Maximum Usage the only routes are waiting or a pay-as-you-go API key.

Plan or pay per token?

For scale, 10M input and 1M output tokens on Muse Wonder cost $60 at list price, or $25.80 if 9M of that input hits the cache. Compare that with the $15 or $50 plans before you upgrade. Full rates are on Meta API prices, and Muse Wonder vs GPT-6.1 Sol shows how it stacks up on price and LMArena scores.

Hit the limit in the Meta Muse app instead?

The Muse app has its own plans. Free has a usage limit that refreshes (Meta does not publish the window); you can upgrade or wait. Power ($20) gives 500M Muse tokens per week and Maximum ($100) gives 3B Muse tokens per week. See Meta Muse limits explained and is Meta Muse worth it.