Baseten lists Base Labs' elephant-1 and mammoth-1 with no prices yet
Baseten added two models, elephant-1 and mammoth-1 from Base Labs, to the top of its Model APIs catalog.
Baseten’s pricing page · Change Detected
WHY IT MATTERS
For buyers comparing inference providers, an unpriced model cannot be budgeted, so these two rows are a stub until the dashes turn into rates. The thing to watch is whether Baseten fills in the Input, Cache Input and Output columns on its next capture, and whether the rates land near the $0.40 input it just set for Reflection Beam or closer to the pricier GLM rows. A long gap between listing and pricing would make this a pattern worth tracking.
BY THE NUMBERS
- Input, Cache Input and Output show a dash for both new models
- 8 real changes in the last 90 days; average gap about 13 days
- Previous change: Oct 21, 2026, Reflection Beam rates cut up to 72%
- Feature additions: 432 changes across 353 companies in the last 180 days
THEIR RATIONALE
Baseten has not said why the models are listed before they are priced. A live row with a Try Model API button and no rates suggests the models are open for testing while pricing is still being set, which is an inference, not something the page states. The timing fits the company's recent pace: it has logged 8 real changes in the last 90 days, an average gap of about 13 days, and this one lands a day after the Oct 21 Reflection Beam rate cut.
THE SIGNAL
Catalog additions are routine for Baseten and for the index: 432 pricing changes across 353 companies involved feature additions in the last 180 days, 21.5% of the 2009 changes tracked. What stands out is the missing price. Earlier additions in Baseten's own history carried rates on day one: Kimi K3 on Jul 28 at $3.00 per 1M input tokens, and DeepSeek V4.1 Flash 1017 on Oct 18 at $0.30 input and $1.20 output per 1M tokens. This time the models are listed with a Try Model API button before any rate is published.
IN CONTEXT
Baseten has been adding to its catalog every few days: the Reflection Beam rate cut on Oct 21 came a day before this listing, and DeepSeek V4.1 Flash 1017 on Oct 18 and Reflection Beam and Meta Muse Spark Open on Oct 17 landed just before that. Those launches carried prices, such as DeepSeek V4.1 Flash 1017 at $0.30 input and $1.20 output per 1M tokens, while elephant-1 and mammoth-1 do not. Peers usually list new models with rates attached, as fal did with its FLUX 3 Max video model at $0.085/sec on Oct 20. The next capture will show whether Baseten fills in the rates.