Fangcun AI publishes third-party safety evaluation of DeepSeek V4.1 Flash-1017 for manipulation risks
pstAsiatech · x · 2026-10-28
Fangcun AI released an independent third-party safety evaluation of DeepSeek V4.1 Flash-1017, focusing on loss-of-control and harmful manipulation risks. The report benchmarks it against DeepSeek V4.1/V4 Pro/V4 Flash, GPT-5.2 Arc, GPT-5.2 Se, and Gemini 3.0 Flash. The full report is publicly available at fangcunleap.com.
More from Models
- Liquid AI launches LFM3: tri-hybrid architecture, six edge models, all modalities — maximelabonne · 2026-10-28
- Google's Genie 4 debuts: real-time playable world model with native audio and live prompt edits — _philschmid · 2026-10-28
- Stanford CS312 quiz: scale weight decay 8x when batch size grows 8x — tatsu_hashimoto · 2026-10-28
- Matthew Berman: DeepSeek v4.1 flash is the best open-weights model right now — MatthewBerman · 2026-10-28
- Google exec teases Genie 4, its new frontier world model, coming to the API — OfficialLoganK · 2026-10-28
- Microsoft's MAI Cyber model doubles user adoption to 50% in 30 days — mustafasuleyman · 2026-10-28