The Institute of Foundation Models has released K2 Horizon, a family of six models ranging from 0.9 billion to 375 billion parameters.
The unusual part is the depth of the release. IFM is publishing final weights and code under Apache 2.0, together with intermediate checkpoints, training configurations, evaluation results, logs and training data where licensing permits. Where the underlying data cannot be redistributed, it says it provides construction methods and mixture recipes.
The family spans very different hardware budgets: smaller variants target constrained and local devices, while the 32B, 36B-A4B and 375B-A23B models are aimed at progressively heavier local, on-premise and enterprise workloads.
IFM reports strong results across reasoning, mathematics, coding and agentic tasks, but those benchmark claims still need broader independent evaluation. The release is notable even before that happens because it gives developers substantially more visibility into how the models were built.