Skip to content
PricingSaaS
Sign in
Back
Pricing briefing
MMindsDB
MindsDB
PACKAGING CHANGE

MindsDB added Muse Ember to its model rate card

By , Co-founder, PricingSaaS

Published

MindsDB added Meta's Muse Ember 1.0 model to its token pricing table.

See before and after

MindsDB’s pricing page · Change detected

What changed

MindsDB added Meta's Muse Ember 1.0 to its token-based rate card at $0.10 input and $0.50 output per 1M tokens, per MindsDB's pricing page as Pulse saved it on Nov 11, 2026. The same page now also carries a new “Decision and embedding models” table, per the same Nov 11, 2026 page: Text Embedding 3 (small) priced at $0.02 per 1M tokens (embeddings carry input-only pricing on the page), and Jev 1.13.0 at $0.

This new model lands at the identical $0.10 input / $0.50 output rate MindsDB set for Claude Haiku 5.5 on Nov 7, 2026, per Pulse's tracked history — just four days earlier.

Why it matters

For a product leader selling AI features, MindsDB's approach is instructive: rather than reprice seat tiers, it runs a pure usage-based rate card across many third-party models and simply adds a row whenever another model launches. Per Pulse's tracked history, MindsDB logged 18 pricing-page changes in the last 90 days, roughly one every 9 days — closer to a SKU catalog that never stops growing than a typical SaaS price list. This addition arrives at the same $0.10 input / $0.50 output rate MindsDB set for Claude Haiku 5.5 on Nov 7, 2026, per Pulse's tracked history, a sign that pricing for small, cheap models is compressing toward a shared floor.

Teams charging a flat per-seat or flat-credit price for AI features built on these same underlying models should watch whether that floor keeps falling — every drop in the models they resell either gets passed to customers as margin erosion or pocketed as margin expansion, and MindsDB's cadence shows how fast that catalog can move.

In context

Per Pulse's tracked history, MindsDB has made 29 real pricing changes to this page since tracking began on Mar 9, 2026, most recently when it added Claude Haiku 5.5 on Nov 7, 2026 at the identical $0.10/$0.50 rate the new model now carries. Of the 35 events Pulse has tracked here since Mar 9, 2026, 9 were new-feature additions and only 2 were price cuts, and MindsDB also added Kimi K3.1 Pro and DeepSeek V4.1 Flash on Oct 27, 2026. That sits inside a bigger shift: Pulse tracked 377 credits-and-usage-billing changes across 259 companies in the last 180 days, 12.3% of everything it tracked, with Pika (8) and Lumaai (7) moving fastest. The sharper read is less about this one model and more about MindsDB's rate card becoming a live read of what model providers charge: per Pulse's tracked history, a new row lands roughly every 9 days on average, and this one landing at last week's exact price is the clearest sign yet that cheap small-model pricing is converging.

Common questions

What would this cost a team sending real volume through the new model?

MindsDB bills purely by usage, not seats: take your daily input and output token counts and run them against the per-million rates on its pricing page (as Pulse saved it on Nov 11, 2026) — there's no minimum commitment or seat fee to layer on top.

Is there a way to try the new model before committing to paid usage?

Yes — MindsDB's Free tier, part of the Free / Pay-as-you-go split it moved to on Oct 12, 2026, covers every model in the catalog billed at provider rates with no monthly minimum, so testing a new listing costs nothing until you actually call it.

Does MindsDB's usage pricing come with promos that could disappear later?

It has before: per Pulse's tracked history, MindsDB ended a 100% credit match promo worth up to $5,000 for new teams on Oct 3, 2026 — sign-up incentives here have come and gone without a broader plan change.

Sources