K2 Horizon 7B

K2 Horizon 7B is the medium dense member of IFM's (MBZUAI) fully open K2 Horizon family: a 7B-core decoder-only reasoning model (36 layers; about 9.0B parameters in the released safetensors including its 250K-token vocabulary embeddings) pretrained on 22.9T tokens, with a native 524,288-token (512K) context window from midtraining onward. Text in, text out, with per-request reasoning effort (high recommended) and JSON/XML tool-call formats. Released 2026-09-03 under Apache 2.0; on 2026-10-01 the main branch was updated to an anti-doom-loop (FTPO) checkpoint that cuts runaway repetitive reasoning. BF16 weights plus FP8 and GGUF builds, an optional Uno discrete-diffusion LoRA adapter for faster decoding, intermediate checkpoints, training data, code, and logs. No priced provider is cataloged yet.