Agent Arena更新了模型评估功能，Claude Opus 5在代码任务中表现出色，GitHub集成让编码更便捷。
Agent Arena在数百万真实世界、长期目标型任务上对模型进行评估，Claude Opus 5在代码任务中排名第一。GitHub集成提升编码能力，支持深度仓库集成，让用户在浏览器中完成编码任务。
Agent Arena measures models on millions of real-world, long-horizon agentic tasks: Overall, and spec...
Agent Arena measures models on millions of real-world, long-horizon agentic tasks: Overall, and specifically by Code, Chat or Work. The additional coding improvements through our GitHub integration will contribute to powering our leaderboard! In Agent Arena: Code, the top 15 rankings are currently: #1 Claude Opus 5 (Max) by @AnthropicAI +14.12% #2 Claude Opus 5 (High) by @AnthropicAI +13.29% #3 Claude Fable 5 (High) by @AnthropicAI +12.82% #4 GPT 5.6 Sol (xHigh) by @OpenAI +12.77% #5 Claude Opus 4.8 (High) by @AnthropicAI +11.78% #6 Kimi K3 (Max) by @Kimi_Moonshot +11.19% #7 Claude Sonnet 5 (High) by @AnthropicAI +11.05% #8 DeepSeek V4 Pro (High) (0813) by @deepseek_ai +11.04% #9 GPT 5.5 (High) by @OpenAI +8.79% #10 GPT 5.5 (xHigh) by @OpenAI +8.54% #11 Claude Opus 4.7 (High) by @AnthropicAI +8.16% #12 Claude Opus 4.7 by @AnthropicAI +7.69% #13 GLM 5.2 (Max) by @Zai_org +7.25% #14 GPT 5.5 by @OpenAI +6.94% #15 Grok 4.5 by @SpaceXAI +6.50% Arena.ai @arena Coding in Agent Mode with GitHub is now available on @arena ! We’ve overhauled our Agent Mode architecture to support deep repository integration. By integrating directly with GitHub, Agent Mode lets our users connect a repo, complete a coding task, and push changes without leaving their browser. AI coding assistants have changed how code is created, but for many developers, chat has still been a “read-only” or “copy-paste” experience. Until now. See more features included in this update below. Your browser does not support the video tag. 🔗 View on Twitter 🔗 View Quoted Tweet 💬 12 🔄 2 ❤️ 54 👀 7970 📊 13 ⚡