MindsDB adds two more models to MindsHub rate card, hours after its last one
MindsDB added two more models to the MindsHub token-based rate card: MindsHub Blaze, a Cerebras-hosted model priced at $0.35 per 1M input tokens / $0.75 output / $0.35 cached, with free cache writes; and GPT-OSS 120B via Fireworks AI at $0.15 input / $0.60 output / $0.02 cached / $0.15 cache-write per 1M tokens.
MindsDB’s pricing page · Change Detected
WHY IT MATTERS
As model-router platforms compete on catalog breadth as much as margin, rate cards start behaving like product listings that get refreshed continuously instead of pricing pages that change quarterly. Watch whether MindsDB keeps this near-daily pace through the rest of September or whether it was a one-day catch-up batch, and whether rival routers respond by matching update frequency rather than just price.
BY THE NUMBERS
- 18 real pricing changes in the last 12 months; 13 in the last 90 days
- Second rate-card update the same day (previous: hours earlier, Sep 14)
- 584 credits/usage-billing changes across 329 companies in 180 days (10.8% of all tracked)
- GPT-OSS 120B priced at $0.15/1M input, among the cheapest models on the card
THEIR RATIONALE
MindsDB isn't touching plan structure or price levels here, just widening model choice on a rate card it now updates constantly. This is its 18th real pricing change since tracking began in March and its 13th in the last 90 days alone — a roughly 11-day average cadence that this same-day repeat blows past. The framing is breadth: give Cowork/Inference API users a cheap open-weights option (GPT-OSS 120B) alongside a fast proprietary one (Blaze) without waiting for the next scheduled update.
THE SIGNAL
Two rate-card pushes in one day is a tell that model additions have become routine maintenance rather than news events for MindsDB — its seventh touch on credits/usage billing since June's MindsHub rebrand. That fits a broader shift: 584 pricing changes across 329 companies in the last 180 days involved credits or usage-based billing, about 1 in 9 of everything the index tracked. GPT-OSS 120B's $0.15/1M input price also puts it among the cheapest entries on the card, alongside frontier-priced options — MindsDB is bracketing its catalog rather than picking a lane.
IN CONTEXT
This is MindsDB's second rate-card expansion of the day, following four models added just hours earlier, including GPT-6 Astra and Claude Fable 5.1. It's the seventh time since June's rebrand to MindsHub that the company has touched its credits-and-usage setup, part of a run that also includes moving Pro billing to a prepaid balance and cutting per-token prices ~30% across 17 models. The same usage-billing pattern showed up elsewhere today too, at Casemark and Coderpad — a sign that continuous, model-priced rate cards are becoming the norm rather than the exception across AI tooling.