MEGA Hub

Evaluating DeepSeek V4.1 Flash-1017 for loss of control and harmful manipulation risks

Authors

Do you know Yao Huang?You can claim authorship or link another user.Do you know Zhiyuan Zeng?You can claim authorship or link another user.Do you know Yu Wang?You can claim authorship or link another user.Do you know Zhitong Zhu?You can claim authorship or link another user.Do you know Dingcheng Duan?You can claim authorship or link another user.Do you know Hao Li?You can claim authorship or link another user.Do you know Shangzhi Xu?You can claim authorship or link another user.Do you know Yifan Wang?You can claim authorship or link another user.Do you know Jinzhe Tu?You can claim authorship or link another user.Do you know Shukun Xiong?You can claim authorship or link another user.Do you know Yihe Zang?You can claim authorship or link another user.Do you know Han Yan?You can claim authorship or link another user.Do you know Hongchang Yang?You can claim authorship or link another user.Do you know Fuyu Gu?You can claim authorship or link another user.Do you know Wei Xu?You can claim authorship or link another user.Do you know Yinpeng Dong?You can claim authorship or link another user.Do you know Xiaojian Li?You can claim authorship or link another user.

Abstract

As frontier language models grow increasingly capable and autonomous, independent third-party evaluations are critical for characterizing their systemic safety risks. While closed-source releases are routinely accompanied by comprehensive system cards, open-weight models receive less scrutiny regarding frontier risks. To address this gap, we present an independent safety evaluation of DeepSeek V4.1 Flash-1017, the latest version of V4.1 Flash that departs from V4 with a new architecture, scaled-up training, and strong agentic capabilities. Drawing on the EU General-Purpose AI Code of Practice, we evaluate the model across 34 benchmarks, including 5 in-house suites, spanning evaluation integrity, loss of control, and harmful manipulation. Conforming to the AEF-1 standard, we compare V4.1 Flash-1017 against V4.1 Flash, earlier DeepSeek V4 models, and leading closed-source systems (GPT-6 Astra, GPT-5.6 Sol, and Gemini 3.8 Flash). Overall, V4.1 Flash-1017 shows substantial safety improvements over the V4 generation and approaches or matches leading closed-source baselines on several dimensions, while still exhibiting notable behaviors that warrant continued attention. Building on our findings, we highlight three key observations and recommendations. First, because safety can shift and even regress across minor updates, developers should integrate safety evaluation into incremental releases. Second, as final-turn outputs and end-state checks alone cannot reliably detect loss-of-control behaviors, deployments require process-level observability across both reasoning traces and action logs. Third, because public benchmarks elicit strong evaluation awareness and face data contamination risks, developing private benchmarks is essential for trustworthy frontier evaluation.

Community

00