微软 MAI-Code-1-Flash 在仅有 5B 激活参数的情况下，SWE-Bench Pro 得分达 51%

微软 MAI-Code-1-Flash 在仅有 5B 激活参数的情况下，SWE-Bench Pro 得分达 51%
Microsoft's MAI-Code-1-Flash Scores 51% SWE-Bench Pro with Just 5B Active Params

原始链接: https://microsoft.ai/models/mai-code-1-flash/

专为解决最关键的复杂问题而打造，兼具卓越的推理能力与顶尖的 SWE-Bench Pro 测试成绩，且价格适中。内置于 GitHub Copilot 和 VS Code 中的轻量级智能代理模型，助力您的工程团队更高效地编写优质代码。根据任何文本或图片提示，创作出精美且可直接用于设计的图像，并以更低的价格获得同类领先的 Arena ELO 评分。将嘈杂的音频转化为精准的行业领域转录文本，并拥有领先的 FLEURS 和 Artificial Analysis 准确度得分。

Built for the complex problems that matter most, with competitive reasoning and top SWE-Bench Pro results at a mid-weight price.

Helps your engineering team write better code faster, with a lightweight, agentic model built into GitHub Copilot and VS Code.

Create stunning, design-ready images from any text or photo prompt, with best-in-class Arena ELO scores at a lower price.

Turn noisy audio into precise, domain-specific transcripts, with leading FLEURS and Artificial Analysis accuracy scores.

微软 MAI-Code-1-Flash 在仅有 5B 激活参数的情况下，SWE-Bench Pro 得分达 51% Microsoft's MAI-Code-1-Flash Scores 51% SWE-Bench Pro with Just 5B Active Params

微软 MAI-Code-1-Flash 在仅有 5B 激活参数的情况下，SWE-Bench Pro 得分达 51%
Microsoft's MAI-Code-1-Flash Scores 51% SWE-Bench Pro with Just 5B Active Params