# SafeWork-R1: Coevolving Safety and Intelligence under the AI-45circ Law

arXiv: https://arxiv.org/abs/2507.18576 · HTML: https://mlsanity.com/paper/2507.18576/

## Key facts

- Cited by 14 papers in the map (data through 14 Oct 2026).
- Placed in Reasoning for agents, planning and code, one of 17 directions in LLM reasoning.
- First posted on arXiv 24 Jul 2025.

## Placed in

- LLM reasoning / [Reasoning for agents, planning and code](https://mlsanity.com/lr/reasoning-for-agents-planning-and-code/)

## Built on by (most cited first, 14 of 14)

- [A Survey of Scientific Large Language Models: From Data Foundations to Agent Frontiers](https://mlsanity.com/paper/2508.21148/) (2025)
- [OpenRT: An Open-Source Red Teaming Framework for Multimodal LLMs](https://mlsanity.com/paper/2601.01592/) (2026)
- [Refusal Falls off a Cliff: How Safety Alignment Fails in Reasoning?](https://mlsanity.com/paper/2510.06036/) (2025)
- [Reasoning over Boundaries: Enhancing Specification Alignment via Test-time Deliberation](https://mlsanity.com/paper/2509.14760/) (2025)
- [EvoDefense: Co-Evolving Black-Box Defense with Large Language Models](https://mlsanity.com/paper/2605.31140/) (2026)
- [Conditional Advantage Estimation for Reinforcement Learning in Large Reasoning Models](https://mlsanity.com/paper/2509.23962/) (2025)
- [Contextual Image Attack: How Visual Context Exposes Multimodal Safety Vulnerabilities](https://mlsanity.com/paper/2512.02973/) (2025)
- [Think-Reflect-Revise: A Policy-Guided Reflective Framework for Safety Alignment in Large Vision Language Models](https://mlsanity.com/paper/2512.07141/) (2025)
- [Spider-Sense: Intrinsic Risk Sensing for Efficient Agent Defense with Hierarchical Adaptive Screening](https://mlsanity.com/paper/2602.05386/) (2026)
- [ADORA: Training Reasoning Models with Dynamic Advantage Estimation on Reinforcement Learning](https://mlsanity.com/paper/2602.10019/) (2026)
- [DAIN: Dynamic Agent-Based Interaction Network for Efficient and Collaborative Multimodal Reasoning](https://mlsanity.com/paper/2606.30189/) (2026)
- [Deep Interaction: An Efficient Human-AI Interaction Method for Large Reasoning Models](https://mlsanity.com/paper/2607.14049/) (2026)
- [SafeRI: Recognition and Intervention for Token-Level Safety Intervention in Large Vision Language Models](https://mlsanity.com/paper/2609.03544/) (2026)
- [ReSI: Recursive Safety Improvement toward Resistant and Resilient AI](https://mlsanity.com/paper/2610.12233/) (2026)

## Abstract (arXiv)

We introduce SafeWork-R1, a cutting-edge multimodal reasoning model that demonstrates the coevolution of capabilities and safety. It is developed by our proposed SafeLadder framework, which incorporates large-scale, progressive, safety-oriented reinforcement learning post-training, supported by a suite of multi-principled verifiers. Unlike previous alignment methods such as RLHF that simply learn human preferences, SafeLadder enables SafeWork-R1 to develop intrinsic safety reasoning and self-reflection abilities, giving rise to safety `aha' moments. Notably, SafeWork-R1 achieves an average improvement of 46.54% over its base model Qwen2.5-VL-72B on safety-related benchmarks without compromising general capabilities, and delivers state-of-the-art safety performance compared to leading proprietary models such as GPT-4.1 and Claude Opus 4. To further bolster its reliability, we implement two distinct inference-time intervention methods and a deliberative search mechanism, enforcing step-level verification. Finally, we further develop SafeWork-R1-InternVL3-78B, SafeWork-R1-DeepSeek-70B, and SafeWork-R1-Qwen2.5VL-7B. All resulting models demonstrate that safety and capability can co-evolve synergistically, highlighting the generalizability of our framework in building robust, reliable, and trustworthy general-purpose AI.