OpenRouter
PulseAugur coverage of OpenRouter — every cluster mentioning OpenRouter across labs, papers, and developer communities, ranked by signal.
16 天有情绪数据
-
Tencent's Hunyuan Hy3 leads API calls; Juyue Xingchen releases Step Image Edit 2
Tencent's Hunyuan Hy3 preview model has topped OpenRouter's latest global API call volume rankings, also securing first place in tool usage and second in programming tasks. Separately, Jieyue Xingchen has launched its n…
-
AI模型在复杂基准测试中接受测试;DeepSeek 4 Pro服务器熔化
一位用户正尝试对DeepSeek 4 Pro模型进行基准测试,但其服务器正经历高负载。该基准测试涉及一项复杂的逆向工程任务,旨在创建一个用于构建Apollo GraphQL哈希的工具。到目前为止,没有开源模型成功完成该基准测试,而Anthropic的Opus 4.7和OpenAI的GPT 5.5等专有模型已显示出成功。
-
Poolside AI 发布开源 Laguna XS.2 和 M.1 编码模型
Poolside AI 发布了两款新的代理式编码模型 Laguna M.1 和 Laguna XS.2,以及它们的代理训练和运行时间。Laguna M.1 是一个大型混合专家(MoE)模型,在 NVIDIA Hopper GPU 上使用 30T 个 token 进行训练,而 Laguna XS.2 是一个较小的开源模型,可在 Apache 2.0 许可下使用。这些模型专为长周期任务设计,旨在实现能够编写和执行代码的更强大的 AI 代理。
-
Qwen 3.6 Plus outperforms DeepSeek V4 Pro in price and quality benchmarks
A recent battle test of six April-released Large Language Models (LLMs) revealed that the Qwen 3.6 Plus, released 22 days prior, outperformed the newer DeepSeek V4 Pro. Despite DeepSeek V4 Pro's advanced reasoning archi…
-
OpenKB 和 OpenRouter 实现无向量AI知识库;揭示LoRA的生产限制
一项新研究表明,LoRA和QLoRA微调方法所依据的低秩假设在生产环境中可能不成立。虽然这些技术能够在有限的硬件上高效地适应大型语言模型,但实际应用经常违反均匀分布的假设,导致性能问题。这一发现可能会对定制化LLM的开发和部署产生重大影响。
-
Eden AI offers unified API for diverse AI models with smart routing
Eden AI has launched a new platform designed to simplify access to various artificial intelligence models through a single API. The service aims to provide developers with a unified interface for integrating leading Lar…
-
Reka AI's Edge model now available on OpenRouter platform
Reka AI has made its Reka Edge model accessible through the OpenRouter platform, enabling developers to integrate its multimodal capabilities without managing their own infrastructure. This integration allows users to l…
-
Google DeepMind launches autonomous research agents powered by Gemini 3.1 Pro
Google DeepMind has launched two new autonomous research agents, Deep Research and Deep Research Max, powered by Gemini 3.1 Pro. These agents are designed to securely analyze user-provided or third-party data, with Deep…
-
Moonshot Kimi K2.5 - 以一半的成本击败 Sonnet 4.5,SOTA开源模型,首个原生图像+视频,100个并行Agent Swarm管理器
Moonshot 发布了 Kimi K2.6,一个更新的开源模型,增强了其在代理编码和多模态理解方面的能力。新版本采用 1T 参数的混合专家架构,拥有 32B 激活参数和 384 个专家,支持 256K 上下文窗口以及原生图像/视频处理。Kimi K2.6 在各种编码和推理基准测试中声称达到了最先进的性能,包括具有数千次工具调用和扩展自主运行的长视距任务。
-
BrowserOS 发布开源浏览器,内置 AI 代理实现网页自动化
BrowserOS 已作为一款专为 AI 时代设计的开源浏览器发布,集成了 AI 代理,可通过自然语言命令自动执行网页任务。它优先考虑用户隐私,并通过支持 Anthropic Claude、Google Gemini 和 OpenAI 等热门选项以及本地模型在内的 11 家以上 AI 提供商来提供广泛的自定义。该浏览器基于 Chromium 分支构建,可确保与现有 Chrome 扩展程序的兼容性,并为普通用户和开发人员提供用户友好的体验。
-
Anthropic upgrades Claude Sonnet, Cursor valued at $28B
Anthropic has released an upgraded version of its Claude 3.5 Sonnet model, which reportedly matches the capabilities of its Opus 4.6 counterpart in some benchmarks and offers a 1 million token context window. Independen…
-
Google 推出代理记忆框架;DeepSeek 发布经济高效的 V4 模型
Google Research 推出了 ReasoningBank,这是一个新颖的框架,旨在增强 AI 代理在部署后从成功和失败的经验中学习的能力。该系统从过去的交互中提炼出可泛化的推理策略,使代理能够持续改进并避免重复错误。另外,新的研究探索了通过潜在表示优化多代理通信,并为在开放式环境中运行的代理引入了 Agent Evolving Learning (AEL),重点关注如何有效利用记忆信息。此外,DeepSeek 发布了其 V4…