@andrei_chiffa@mastodon.social
Post #4060610
2026-07-24 10:15 UTC
Free, fully open, multilingual, copyrights-compliant, Swiss LLM just expanded!
@apertus@fosstodon.org 1.5 family is out, with tools, reasoning, and image/audio multimodality, and significant base improvement across the tasks in text2text mode.
It's really good darn good, and we also went out of our way to make sure it's safe: https://apertus-ai.org/pages/get-started/
1/
Replies (1)
-
@andrei_chiffa@mastodon.social 2026-07-24 10:17
We confirmed the model wasn't just refusing to provide dangerous information, but was unable of it, by performing abliteration ourselves (worst-case jailbreak; by Arthur Wuhrmann based on his experience at Surelio.ai). 2/