Tavus Growth: 1,250 minutes for $397 vs a per-clip Sume budget

Tavus's pricing page lists Growth at $397 a month for 1,250 conversational video minutes, 7 custom replicas and up to 10 concurrent streams, and Starter at $59 for 100 minutes. Overage is listed at $0.37 a minute on Starter and $0.32 on Growth. Divide and you get about $0.32 a minute on Growth and $0.59 on Starter when the bucket is fully used. Sume has no minute bucket: each avatar clip or Format run is its own paid job, and a generation_spend_cap_usd on a Format run (up to $500) sets the ceiling before it starts.

Tavus source: pricing, read 2026-10-04. Sume sources: Calling a Format and Jobs and results.

Effective rates, from the page

The effective rate holds only if you use the whole bucket. Unused minutes raise the real price per minute, and the page lists overage on Starter, so a busy month costs more than the base.

Why the comparison is hard

A minute of live conversation depends on how long people talk, which you do not control. A rendered clip has a known length before you submit it, so the cost is a function of script length and tier. That makes a Sume budget something you can compute in advance, and the spend cap makes it something you can enforce.

• Live minutes: demand-driven, bucketed by month.
• Rendered seconds: known upfront, billed per job.
• One rendered clip can reach many viewers; a live minute reaches one.

Making the Sume budget concrete

Take the number of clips you plan, multiply by the per-clip price for your tier from the current pricing page, and add a margin for retakes. Set the run cap above one run's expected spend but well below the monthly total. Each receipt reports what the run actually spent as usage.billable_amount_usd_micros, so you can reconcile against the cap.

Bottom line

The two plans measure different units. Decide if the product is a conversation or a clip, then compare the unit that matches. Use your own usage forecast and the vendor's current pricing, not this post's arithmetic, to decide.