Baseten Releases Joey-1 Decision
The interesting bet here is that a huge share of AI workloads are really just classification — routing, moderation, risk scoring — where generating text is wasted effort. If joey-1's calibrated probabilities hold up in production, 20 ms latency and $0.01/M tokens make LLM-based classification look overpriced.… more
But the real competitor isn't a frontier chat model, it's XGBoost and fine-tuned encoders: free, battle-tested, and already fast. The 'calibrated' claim is also the fragile part — probabilities drift under distribution shift, and launch-day calibration says little about month six. Still, an inference provider shipping its own model hints at where margins are heading: specialized small models, not bigger chatbots.
Baseten Releases joey-1, a 2B Decision Model at $0.01 per 1M Tokens
Baseten has released joey-1, a 2B open-weight decision model that returns calibrated probabilities instead of text. It runs at 20 ms P50 latency and costs $0.01 per million input tokens. The post Baseten Releases joey-1, a 2B Decision Model at $0.01 per 1M Tokens appeared first on…
Source:
Marktechpost
Discussion (0)