公司
别名：@harvey
法律科技AI公司，开发内部后台Agent Spectre用于工程协作。
已跟踪 11 条高相关材料
最近变化
2026-06-03 · Harvey将内部Agent Spectre集成到Devin Desktop，打通工具链上下文。
为什么值得关注
Harvey 被反复提及时，通常意味着它正在影响产品路线、开发者工作流或 AI 产业判断。这个页面把分散材料合并成一个可持续更新的观察入口。
Frontier models are powerful advisors. On @harvey's Legal Agent Benchmark, a GLM 5.1 worker using C...
Fireworks AI(@FireworksAI_HQ) · 8.7 分
Fireworks AI 通过“harness + advisor”架构，在 Harvey 法务代理基准上以 Claude Opus 4.7 为稀疏顾问，将 GLM 5.1 工作者性能提升至 18/100 全对，成本仅为 Opus 39%。
https://t.co/UnoEzV2tZs
Harrison Chase(@hwchase17) · 8.5 分
LangChain Labs启动，专注于持续学习研究，推动智能体自我改进技术发展。
Building with Gemini Embedding 2: Agentic multimodal RAG and beyond
Google Developers Blog · 8.5 分
Google宣布Gemini Embedding 2正式可用，该模型支持文本、图像、视频、音频和文档的统一嵌入，实现100多种语言的跨模态搜索与应用，如增强型多模态RAG、视觉搜索等。
已收录 11 条与 Harvey 相关的内容，按评分排序。
Fireworks AI demonstrates that GLM 5.1, when using Claude Opus 4.7 as a sparse advisor in the Legal Agent Benchmark, achieves 18/100 all-pass versus 14/100 for Opus alone at 39% of the cost.
入选理由：在 Harvey 法务代理基准上，GLM 5.1 + Claude Opus 4.7 稀疏顾问方案全对数达 18/100。
LangChain Labs launches, focusing on continual learning research to advance self-improving agent technology.
入选理由：LangChain Labs聚焦持续学习，提升智能体自我优化能力
Google宣布Gemini Embedding 2正式可用，该模型支持文本、图像、视频、音频和文档的统一嵌入，实现100多种语言的跨模态搜索与应用，如增强型多模态RAG、视觉搜索等。
入选理由：Gemini Embedding 2是首个通过单一接口处理多样输入并映射至同一语义空间的模型，支持多模态数据理解。
Routing and post-training open-source models significantly improve AI system accuracy, speed, and cost-efficiency. Harvey and Fireworks AI demonstrated that a hybrid architecture using GLM 5.1 as the primary worker with selective frontier model routing achieves superior quality and lower costs in legal tasks, proving this approach is a viable alternative to pure frontier models.
入选理由：Harvey实测显示混合法律Agent在质量和成本上均优于单一前沿模型。
Fireworks AI and Harvey team jointly launched the Legal Agent Benchmark to evaluate how open-weight models perform on long-horizon legal tasks.
入选理由：Legal Agent Benchmark 是首个聚焦长期法律任务的开源评估基准，支持 Open-Weight 模型测试。
Harvey's engineering team integrated their internal background agent Spectre into Devin Desktop, enabling organizational context to reside on engineers' laptops and flow across AI agents.
入选理由：Harvey将内部Agent Spectre集成到Devin Desktop，打通工具链上下文。
Harvey's coding agents are designed to be collaborative, where engineers, product managers, and researchers all contribute context to the same agent, rather than running parallel ones on separate laptops.
入选理由：Harvey的编码代理通过协作设计提高了团队效率，减少了重复工作。
LangChain announces research collaboration with Harvey on LAB, an open-source benchmark for evaluating AI agents in long-horizon legal tasks.
入选理由：LangChain与Harvey合作开发Legal Agent Benchmark（LAB）。
Harvey's DAU/MAU ratio exceeds 50%, indicating a significant increase in user activity.
入选理由：Harvey 的 DAU/MAU 达到 50%
The article shares user data for Harvey, showing activity over 50%.
入选理由：Harvey DAU/MAU 超过 50%
Harrison Chase shares Harvey's new open-source long-horizon legal agent benchmark, calling for better evaluation frameworks for AI agents in specialized domains.
入选理由：AI代理在法律领域的应用需要专门的长周期任务基准测试。
