Hugging Face announced that the Apertus 1.5 models are now available. The update expands the 8‑billion and 70‑billion parameter models with image and audio understanding, an optional thinking mode for reasoning, and a context window of 262,144 tokens – four times larger than the original. The new weights were uploaded on July 24 2026 under an Apache 2.0 license and remain fully open.
Why it matters
Developers can now build apps that handle text, images and speech together without proprietary APIs.
