Skip to content
View original post on X: BAAIOfficial
Published
52/100

BAAI releases AREX-2, a 27B model built on Qwen3.8-27B

AIBAAI, with USTC and RUC, releases AREX-2, a 27B model built on Qwen3.8-27B. The post says it scores 81.8 on MLE-bench Lite and sets new ≤40B state-of-the-art results on BrowseComp (84.0) and HLE (52.6). The benchmark chart also reports 93.8 on DeepSearchQA and 92.2 on GAIA.
BAAIVerified on XOfficial
@BAAIBeijing
View on X

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)👇

Source details

BAAI · x.comPublished · added here

Original title: Can a 27B model push the Pareto frontier of autonomous research and beat 1T–2.8T giants?

Topics

AI agentsAI codingReasoningChina AIModel releasesEvals & benchmarks