Baseten Releases joey-1, a 2B Decision Model
Baseten launches joey-1, a 2B open-weight decision model delivering calibrated probabilities at 20 ms P50 latency and $0.01 per 1M input tokens.
Stock photo for illustration only, not from the actual event
- Baseten releases joey-1, a 2B open-weight decision model
- Returns calibrated probabilities instead of standard text generation
- Achieves a P50 latency of 20 milliseconds
- Priced economically at $0.01 per million input tokens
The artificial intelligence landscape welcomes a new release from Baseten, which has introduced joey-1, a 2B open-weight decision model. Unlike traditional language models that generate text, this model is specifically engineered to return calibrated probabilities for streamlined decision-making processes.
In terms of operational performance, joey-1 operates efficiently with a P50 latency of just 20 milliseconds, making it highly suitable for applications requiring rapid, real-time evaluations. Furthermore, it introduces a cost-effective pricing structure set at 0.01 dollars per million input tokens.
By outputting calibrated probabilities rather than raw text, joey-1 allows downstream software systems to integrate statistical likelihoods directly into automated business logic without requiring extra parsing steps, thereby lowering computational overhead.
Stock photo for illustration only, not from the actual event
The introduction of this lightweight yet capable decision model provides developers and enterprise teams with a viable tool for high-speed, low-cost evaluations, expanding the options available for modern AI infrastructure integration.
Source: MarkTechPost
Found something wrong in this article? Report an issue with this article
Comments
Leave a Comment