AI 精选动态
智能评分 88
Apodex 1.0
AI 推荐理由
公开提供 NYC-256K上下文支持的重量文件和基准测试的极高分数，对深度研究代理的实际应用场景有参考价值核心解读
Apodex 向外界发布了Apodex 1.0，一个以验证中心为核心的深度研究代理系统，其关键数据包括150个子代理协同工作模式、基准测试得分（BrowseComp 90.3, DeepSearchQA 94.4, FrontierScience-Olympiad 87.4）以及使用Apache 2.0许可的35B-A3B等大型权重。该系统通过可审计的证据链增强报告准确性， Wissenschaften-Olympiad领域的SOTA地位显现。
全文
Apodex has released Apodex 1.0, a verification-centric deep research agent that searches the web, synthesizes evidence, and generates reports in which every claim is backed by an auditable chain of evidence.
In heavy-duty mode, Apodex 1.0-H runs an async team of up to 150 sub-agents, with a global verifier checking the assembled evidence before any answer is committed.
Evidence over generation 👀
> **引用原帖 Apodex (@Apodex_AI):**
> Meet 𝗔𝗽𝗼𝗱𝗲𝘅 𝟭.𝟬 🔭 — a heavy-duty agent team for deep research, which sets the SOTA! The team searches the web, reasons over evidence, and writes reports where every claim is backed by an explicit 𝘦𝘷𝘪𝘥𝘦𝘯𝘤𝘦 𝘤𝘩𝘢𝘪𝘯, independently audited before delivery.
> 🌐 https://t.co/pOQAjL92uF
> https://x.com/Apodex_AI/status/2064014790788624398
🚨 AI News | TestingCatalog (@testingcatalog): Apodex 1.0-H is reported as a new state-of-the-art across open and closed systems on the public deep-research suite, scoring 90.3 on BrowseComp, 94.4 on DeepSearchQA, and 87.4 on FrontierScience-Olympiad.
The weights ship open under Apache 2.0: Apodex 1.0-mini at 35B-A3B plus 0.8B, 2B, and 4B checkpoints, each on a 256K context with OpenAI-compatible serving via SGLang or vLLM.
Test the hosted agent or pull the weights below
https://t.co/h0Ohh1mJq1
https://t.co/ufSXmXhJ7q