Quick answer
Perplexity API documents sonar model rpm as Tier 0: 5–50 RPM; upper tiers: up to 100–4,000 RPM. Tier 0 allows 50 RPM for sonar, sonar-pro and sonar-reasoning-pro and 5 RPM for sonar-deep-research; Tier 5 allows 4,000 and 100 RPM; tiers unlock at $50, $250, $500, $1,000 and $5,000 cumulative spend.
Verified Sep 17, 2026Official source
Current limits
| Constraint | Current value | Scope | Verified source |
|---|---|---|---|
| Sonar model RPMTier 0 allows 50 RPM for sonar, sonar-pro and sonar-reasoning-pro and 5 RPM for sonar-deep-research; Tier 5 allows 4,000 and 100 RPM; tiers unlock at $50, $250, $500, $1,000 and $5,000 cumulative spend. | Tier 0: 5–50 RPM; upper tiers: up to 100–4,000 RPM | Model and tier specific | PerplexitySep 17, 2026 |
Why does this limit matter?
Using one Sonar limit for every model can under-provision or overload a worker.
This value is scoped to Model and tier specific; a different plan, runtime, model, endpoint, region, or account can produce a different effective constraint.
What should you check?
- Match the exact model or async endpoint to the current tier table.
- Confirm the exact plan, model, runtime, endpoint, region, and account that serve the failing workload.
- Record the observed value, response headers or configuration, timestamp, and source without logging secrets.
Important caveats
- GET polling endpoints have much higher documented RPM than async job creation.
- Treat the official source and live account configuration as authoritative if they differ from this verified snapshot.
HyperObserve reports the documented platform constraint. Your application, SDK, gateway, provider, region, or account can impose a lower effective limit.
Related