IFM Releases K2 Horizon: Six Apache 2.0 Models From 0.9B to 375B
MBZUAI's IFM released K2 Horizon, six Apache 2.0 models (0.9B-375B) with open training data, code, and checkpoints, claiming the largest fully open-source launch.
The Institute of Foundation Models (IFM), launched by MBZUAI, released K2 Horizon: six Apache 2.0 models (0.9B, 3.7B, 7B, 32B, 36B-A4B, 375B-A23B) shipping with the ~20-trillion-token pretraining corpus, intermediate checkpoints, training code, and logs, which IFM calls the largest fully open-source launch in AI history. The 375B-A23B scores 70.2 on Terminal-Bench 2.1 and 87.3 on GPQA Diamond; the 7B model posts 70.6 on SWE-bench Verified. New techniques include MoVA, which extends MoE routing into attention (36B total, ~4B active), and Uno, a LoRA adapter giving roughly 3x lossless decoding speedup. IFM's own reward-hacking audit re-scored 375B-A23B from 70.2% to 66.9% after flagging 24 of 712 Terminal-Bench trials.
- Six sizes share one architecture and serving stack; day-zero vLLM/SGLang/Ollama support
- MoVA moves MoE sparsity into attention; Uno LoRA gives ~3x decoding speedup
- 375B-A23B: 70.2 Terminal-Bench 2.1, 87.3 GPQA Diamond; 7B: 70.6 SWE-bench Verified
- Self-audit for reward hacking cut 70.2% to 66.9% on Terminal-Bench
From the original article
4 min read at marktechpost.com
Most open model launches release one checkpoint and a benchmark table. The Institute of Foundation Models (IFM) released something wider last week. IFM is the frontier lab launched by MBZUAI in May 2025.
By Asif Razzaq · marktechpost.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.