BAAI 联合 USTC、RUC 发布 27B 模型 AREX-2，主打自我进化能力
BAAI 和中科大、人大合作搞了个 27B 的 AREX-2，用自我进化训练，BrowseComp 拿 84.0，号称能打 1T 以上的巨无霸。
BAAI 联合中科大与人大推出 AREX-2，基于 Qwen3.8-27B 构建，参数量仅 27B。模型将递归自我改进（RSI）落地为测试时扩展范式，重点训练深度反思与长程执行两项能力。成绩上，AREX-2 在 MLE-bench Lite 达 81.8，BrowseComp 得 84.0（≤40B 档新 SOTA），HLE 得 52.6。论文称其复杂编程与深度研究表现可对标 1T-2.8T 参数量的超大模型。
Can a 27B model push the Pareto frontier of autonomous research and beat 1T–2.8T giants?
Meet AREX-2 (27B), spearheaded by BAAI in collaboration with USTC and RUC (built on Qwen3.8-27B). Turning Recursive Self-Improvement (RSI) into a practical test-time scaling paradigm, we pinpoint two keys to self-evolution: Deep Reflection & Long-Horizon Execution. By reinforcing algorithmic optimization tasks, AREX-2 emerges with exceptional deep reflection and long-horizon execution capabilities—significantly boosting its performance oncomplex coding and deep research to top MLE-bench Lite (81.8) and set new ≤40B SOTA on BrowseComp (84.0) & HLE (52.6)👇
