Mapika Decider
Open-weight decision models fine-tuned from Qwen3.5: choices, scores and yes/no probabilities from one forward pass
Decider is a family of open-weight decision models by Mark Marosi (Mapika on GitHub), fine-tuned from Qwen3.5 to reproduce the System One model class that Jev introduced. It reads a state and typed questions and returns, from one forward pass, a probability for every option: a choice over 2 to 255 options, a score over 2 to 10 levels, or a yes/no probability.
The 2B model has a 32K context and a vision variant; a 31B model was added in October. A Python package and a local HTTP server run them.
Apache-2.0 licensed and free to self-host.
Pricing: Free (open source)
Mapika Decider Alternatives
Explore 147 products in the Inference APIs category. View all Mapika Decider alternatives.
Kev
Open decision models from 0.8B to 27B, built on Qwen and compatible with TypeSafe's Jev API.
Laya
Open-source multilingual decision model: typed choices, scores and yes/no answers in one forward pass.
Respan Span-01
Behavior scoring model: the probability that each plain-language behavior you define is present in a conversation
Fastino GLiDE
Fastino's thinking decision model: one selected action, a confidence and a probability for every option
Celeris-1 Decision
Diffusion-based decision model: yes/no, choice and score answers with probabilities and optional one-line reasons
Work on Mapika Decider? Feature it at the top of Inference APIs.
Is your product missing?