Skip to main content
This page previously described a request-priority queue and availability-zone routing for premium accounts. Neither could be confirmed against the Swarms API application source — there is no queue-priority or region-routing logic in the reviewed codebase. Those claims have been removed rather than restated as fact. What follows is limited to premium behavior that is actually implemented and verifiable.

​
What Premium Actually Changes

Premium tiers (pro, ultra, premium, resolved from an active subscription or account tier) get:
  • Higher rate limits: 2,000 requests/minute, 10,000/hour, 100,000/day, versus 100/minute, 350/hour, 1,200/day on the free tier. See Rate Limits.
  • 10x the per-agent token limit: 2,000,000 tokens per agent versus 200,000 on the free tier.
  • Access to 5 premium-only endpoints: batch agent completions, batch swarm completions, reasoning agent completions, batched grid workflows, and graph workflows. See Premium Endpoints.
  • Access to premium-only models: a set of frontier models are rejected on the free tier with suggested free-tier alternatives. See Pricing.
Nothing in the reviewed API source reorders or prioritizes in-flight requests by tier — a premium account gets a materially higher ceiling before hitting a 429, not a faster response for any individual request.

​
Under Heavy Load

If you are being rate-limited, the response carries Retry-After and the X-RateLimit-* headers so you can back off correctly — see Rate Limit Headers. Upgrading raises your rate-limit ceiling; it does not change how the underlying request is scheduled.
To upgrade your plan, visit your Account Dashboard or check the Pricing page.
If you have questions, reach out to support.