Plan limits

How much Muse Spark 1.3 usage do you get on Cursor?

Cursor publishes no token allowance for Muse Spark 1.3. It draws on the Other Models pool, which the Pro and Ultra plans include and the Start plan doesn't, and it bills at $1.25 per million input tokens and $4.25 per million output.

Checked against the lab's own pages on September 17, 2026. Limits change often, so the date matters.

Cursor doesn't give Muse Spark 1.3 a message count or a token allowance, which is what the question usually assumes it has. What you get is a pool of spend.

Cursor's docs split included usage into two pools that reset with your monthly billing cycle. Muse Spark 1.3 is Meta's model rather than Cursor's, so it comes out of the second one.

Input

$1.25 per million input tokens

Output

$4.25 per million output tokens

Cached input

$0.15 per million tokens

Context window

300k tokens, or 1M in Max Mode

Pool

Other Models

Resets

monthly with your billing cycle

What the Other Models pool is

The docs call it the pool for third-party models, charged at the model's API price. Which plans carry it is the part people get caught by.

So the answer to how much Muse Spark 1.3 you get is a division problem, and the divisor is Meta's list price: $1.25 per million input tokens, $4.25 per million output, and $0.15 per million for cached input.

What a long run costs

The context window is 300k tokens, or 1M in Max Mode at the same per-token rates. Cursor's docs say there's no long-context multiplier and no cache-write charge, so a session that fills the window bills at the same rate as a short one (which is not how every model on the list works).

When the pool is empty

Usage resets monthly with your billing cycle, and unused usage does not roll over. Cursor says you'll see a notification in the editor, and from there you either enable on-demand usage, which is pay-as-you-go, or move up a plan.

Two more things to check

Effort first. Muse Spark 1.3 runs at six effort levels, from minimal up to max, and high is the default, so the out-of-the-box setting isn't the cheapest one.

Then access. On Enterprise, an admin has to enable the Meta provider before anyone on the team can pick the model at all.

Cursor publishes the per-token rates on its Models & Pricing page and the pool rules in its usage docs. Neither page prints an included allowance in dollars or tokens, so the arithmetic is yours to do.