Apertus 1.5 releases 8B and 70B models with multimodal understanding and extended context

Apertus 1.5 updates the 8B and 70B Apertus 1.0 models by incorporating image and audio understanding, an optional thinking mode, and a 4x longer context window of 262,144 tokens. The updated weights were released on July 24, 2026, under an Apache 2.0 license. The models were further pretrained using 4 trillion tokens of text and multimodal data for the 8B model, and 2 trillion tokens for the 70B model, totaling 19T and 17T respectively. The models are fully open, featuring open weights, open data, and full training details.