Meta 携 Muse Spark 1.1 加入竞争激烈的 AI 编程战场
英文标题:Meta enters the crowded AI coding battle with Muse Spark 1.1
Meta 于周四公开发布多模态 AI 模型 Muse Spark 1.1,正式进军代理式编码赛道,直接对标 OpenAI 和 Anthropic 的同类产品。该模型定价为输入每百万 token 1.25 美元、输出 4.25 美元,略高于 Claude Haiku 4.5 和 GPT-5.6 Luna,主打处理大规模代理工作负载、缺陷修复与代码迁移等企业自动化场景。CEO 扎克伯格为此打破三年沉默,在 X 平台发帖强调其代理性能与极致性价比,并预告“更多即将到来”。本周 AI 领域迎来密集发布,侧面印证行业竞争持续白热化,开发者与企业正获得更多高性价比的编码 Agent 选择。
62.8 KB
41.9 KB
201.0 KB
中文
Meta 于周四公开发布了新版本 Muse Spark,这是一款专为代理式编码设计的多模态 AI 模型,旨在与 OpenAI 和 Anthropic 提供的类似产品展开竞争。
Spark 1.1 的首个版本于今年 4 月发布,该公司表示,它能够进行多步推理并处理复杂流程、管理数字工作流,以及在 enterprise 系统中部署新功能。
Meta 在这方面略落后于竞争对手;Anthropic 和 OpenAI 提供类似模型已有一段时间。但这并不意味着 Meta 的入局不构成威胁。
AI 行业持续竞争的一个焦点仍然是用量成本,而 Meta 似乎给出了颇具竞争力的费率。据路透社报道,该公司将对每百万输入 token 收取 1.25 美元,对每百万输出 token 收取 4.25 美元。这一价格与 Anthropic 的 Claude Haiku 4.5 和 OpenAI 的 GPT-5.6 Luna 大致持平(尽管略高一点)。
Meta 向用户主打的卖点是 Spark 处理大型代理工作负载、修复缺陷以及辅助大规模代码迁移的能力——而这正是企业越来越多地寻求 AI 公司提供的自动化类型。
“Muse Spark 1.1 在需要跨一系列外部应用与服务进行规划编排的个人代理任务中表现出了卓越性能,”该公司在一篇博文中写道。
Meta 在过去几年里发布了多款基础 AI 模型。Muse Spark 的发布显然重要到足以促使 CEO 马克·扎克伯格三年来首次在 X 上发帖。扎克伯格上一次发帖是在 2023 年 7 月,当时正值该平台从 Twitter 更名为 X。
扎克伯格在帖子中称 Spark 是“一款性能强劲、价格极低的代理和编程模型”,并指出该模型“在代理性能、工具使用和计算机使用方面表现最强”。
扎克伯格还提到“很快还会有更多”——暗示公司计划发布更多模型。
本周对 AI 发布来说意义重大,尤其是对 Meta 而言,该公司还于周二发布了一款名为 Muse Image 的全新 AI 图像生成模型。本周的其他发布还包括 SpaceXAI 的 一个新版 Grok 以及 OpenAI 同样于周四推出的新模型家族 GPT-5.6。可以说,AI 行业的竞争一如既往地激烈,想要脱颖而出的公司面临的任务十分艰巨。
English / 原文
Meta publicly launched a new version of Muse Spark on Thursday, a multimodal AI model designed for agentic coding that aims to compete with similar products offered by OpenAI and Anthropic.
Spark 1.1, the first version of which was announced in April, can engage in multistep reasoning and handle complex processes, manage digital workflows, and deploy new features in enterprise systems, the company says.
Meta is a bit behind its competitors here; Anthropic and OpenAI have offered similar models for quite some time. But that doesn’t mean Meta’s entry into the market isn’t a threat.
An ongoing source of competitiveness within the AI industry remains the cost of usage, and Meta appears to be offering a competitive rate. Reuters reports that the company will charge $1.25 per million input tokens and $4.25 per million output tokens. That puts it in line with (albeit slightly above) Anthropic’s Claude Haiku 4.5 and OpenAI’s GPT-5.6 Luna.
Meta’s pitch to users is Spark’s ability to handle large agentic workloads, fix bugs, and help with large code migrations — the kind of automation that enterprises are increasingly turning to AI companies to provide.
“Muse Spark 1.1 delivers exceptional performance in personal agentic tasks that require planning and orchestration across a range of external apps and services,” the company wrote in a blog post.
Meta has released a handful of foundation AI models over the past few years. The Muse Spark release was apparently important enough to compel CEO Mark Zuckerberg to post on X for the first time in three years. Zuckerberg’s last post was in July 2023, around the time the platform rebranded from Twitter to X.
In his post, Zuckerberg called Spark “a strong agentic and coding model at a very low price,” noting that the model was “strongest at agentic performance, tool use, and computer use.”
Zuckerberg also noted that there was “more to come soon” — implying that the company plans to release additional models.
It’s been a big week for AI announcements — particularly for Meta, which also unveiled a new AI image-generation model on Tuesday, dubbed Muse Image. Other releases this week have included a new version of Grok from SpaceXAI and a new family of models from OpenAI, GPT-5.6, that also dropped Thursday. Suffice it to say that the competition within the AI industry is as healthy as ever, and companies that wish to stand out from their peers have their work cut out for them.