Technology is the core driver of Ant's future creation.
The AntBailing team releases Ring-lite-2507, enhancing deep reasoning through a Two‑staged RL pipeline while simultaneously balancing overall model abilities, showcasing notable gains on benchmarks like ARC‑AGI‑v1 and offering the model as an open‑source resource across major platforms.
The article details the 2025 DCASE Audio Question Answering (AQA) track, outlines its technical challenges, describes Ant Group's three‑stage data, model, and training pipeline, presents performance gains of their Qwen2‑Audio‑R1‑8B and Kimi‑Audio‑SFT‑12B models, and outlines future research directions.
Ant Group’s Ant Financial Science released Agentar‑Fin‑R1, a finance‑focused large model that claims expert‑level knowledge, efficient training, and continuous self‑evolution, outperforming open‑source rivals on benchmarks like FinEval1.0, FinanceIQ and Finova, while supporting industry standards through a collaborative AI alliance.
The newly released “Terminal Agent Security 2025” report, unveiled at the World AI Conference, systematically categorizes AI agent risks, outlines detection and defense methods, and proposes three protection pathways—single‑agent safety, trustworthy multi‑agent interconnection, and AI‑terminal security—to guide the emerging ecosystem of intelligent edge devices.
At the 2025 World AI Conference, Ant Group unveiled an open‑source High‑Order Program framework that blends intelligent models with engineering safeguards to overcome reliability challenges and enable large‑scale, trustworthy AI applications across finance, security, and healthcare.
The CCF‑Ant Research Fund 2025, now open for its first batch, invites global university and institute researchers to apply by August 25 2025 for up to 50 projects spanning data security, hardware‑software co‑design, supercomputing, and artificial intelligence, with detailed topics, eligibility rules, and submission channels provided.
M2-Reasoning-7B, an open‑source 7B multimodal model from Ant Group, combines a high‑quality data pipeline with dynamic multi‑task training and a novel reward function to deliver state‑of‑the‑art performance on both general and spatial reasoning benchmarks, surpassing many larger competitors.
The EXPRESS Workshop at ISSTA 2025, hosted by Ant Group, featured a keynote by Purdue’s Prof. Zhang on an LLM‑driven “Human‑like AI Auditor” called RepoAudit, which demonstrated high‑accuracy automated code review, uncovering dozens of real bugs and hundreds of zero‑day vulnerabilities across major open‑source projects.
The article reviews Yang Xiaofang’s presentation on trustworthy AI, emphasizing the need for systematic support, inclusive design, and participatory governance, and outlines the evolution, capabilities, risks, and multi‑layered solutions for multi‑agent AI systems.
The World Digital Academy unveiled the AI STR series' first global AI Agent Operation Safety Testing Standard, detailing a full‑link risk analysis framework, novel testing methods, and its role in addressing rising safety concerns as AI agents become mainstream in 2025.
