Baseten has unveiled joey-1, a new open-weight decision model featuring 2 billion parameters. This model distinguishes itself by providing calibrated probabilities rather than text outputs, enabling users to make more informed decisions based on its predictions.
With a latency of just 20 milliseconds at the P50 level, joey-1 is designed for efficiency and rapid response. The pricing structure is particularly appealing, set at $0.01 per million input tokens, making it accessible for various applications in the AI landscape.
The launch of joey-1 is significant as it reflects a growing trend in AI towards models that prioritize decision-making capabilities, positioning Baseten as a competitive player in the evolving market of AI solutions.
Compiled automatically by the Tech AI Newsdesk from public AI-news sources and summarised in our own words.
Frequently asked questions
What is joey-1?
joey-1 is a new open-weight decision model launched by Baseten, featuring 2 billion parameters and providing calibrated probabilities.
How fast is the joey-1 model?
joey-1 has a latency of just 20 milliseconds at the P50 level, designed for efficiency and rapid response.
What is the pricing for joey-1?
The pricing structure for joey-1 is set at $0.01 per million input tokens, making it accessible for various applications.
