IFM releases K2 Horizon, six Apache 2.0 models from 0.9B to 375B parameters

The Institute of Foundation Models (IFM), the frontier lab launched by MBZUAI in May 2025, released K2 Horizon last week, a family of six openly licensed language models, MarkTechPost reports. The sizes are 375B-A23B, 36B-A4B, 32B, 7B, 3.7B and 0.9B. All are available on Hugging Face under the Apache 2.0 license, with FP8 and GGUF builds. Unlike a typical single-checkpoint launch, IFM is also publishing the pre-training corpus, intermediate checkpoints, training code, configuration files and fine-grained training logs. The lab calls it the largest fully open-source model launch in AI history. The six models share a core architecture, vocabulary, training methodology, interfaces and deployment tooling, with the exception that the 0.9B model uses a smaller vocabulary. According to the report, the intent is that teams can prototype on the 3.7B model and move up to 375B-A23B without changing their serving stack. Each model was pre-trained on roughly 20 trillion tokens, and nearly 17 percent of the corpus consists of problem-solving trajectories with explicit reasoning. Day-zero support covers vLLM, SGLang and Ollama on NVIDIA, AMD and Cerebras hardware. Hosted APIs are offered through Compass, Cerebras and Nebius via IFM's platform site. The release matters for developers who want a permissively licensed model ladder and for researchers who want full training artifacts rather than weights alone.