隆重推出 Muse Spark 1.3
Introducing Muse Spark 1.3

原始链接: https://research.meta.ai/blog/introducing-muse-spark-1-3

Muse Spark 1.3 现已发布,在智能体任务和编程任务方面实现了显著的性能提升。该模型专为实际应用而设计,现在能够更出色地处理长周期工作流、同时管理多项任务,并能主动修正自身的规划偏差。 主要改进包括: * **增强协作能力**:当面临模糊不清的情况时,模型会提出澄清问题;遇到困难时会请求协助;在执行关键操作前会进行确认。 * **编程效率提升**:在工程任务中,Muse Spark 1.3 更加精简,与前代版本相比,工具调用次数减少了约 20%,Token 使用量减少了约 25%。 * **可靠性增强**:模型展现出更好的约束管理能力,能更准确地执行复杂指令,并对自身局限性有更敏锐的感知,从而减少了幻觉现象。 * **稳健的安全性**:增强了对对抗性输入和提示词注入的防御能力,同时对不可逆操作具有更好的判断力。 Muse Spark 1.3 即日起可通过 Muse Code 和 Meta Model API 使用。目前已支持现有的推理模式,而“最大推理(max reasoning)”功能将在完成最终安全测试后发布。团队计划进行后续更新,包括发布未来的模型版本以及开放权重。

针对“Muse Spark 1.3”(Meta.ai)的 Hacker News 讨论反映出,比起对发布本身感到兴奋,人们更多持有怀疑态度并进行了广泛的行业分析。评论者很快将该贴标记为重复内容,并批评了 Meta 的整体市场定位,一些人甚至认为此次发布不过是其试图“维持存在感”的手段。 讨论的很大一部分集中在人工智能发展的当前轨迹上。用户分析了行业是在 S 型曲线中触及了平台期,还是进展依然处于迭代中。一位贡献者指出,进步往往源于微小而精巧的突破——例如 DeepSeek 对 RLVR 的应用——这些突破即便最终耗尽了潜力,也会暂时改变行业格局。 此外,讨论还涉及了 Meta 的战略挑战,指出该公司拥有庞大的计算资源,在发布之前的模型后进行了内部重组,且在其他领域被视为失败与致力于保持人工智能领域前沿地位之间存在持续的紧张关系。总体而言,舆论持谨慎观察的态度,质疑 Meta 是否真的能在快速演变的市场中保持竞争优势。
相关文章

原文

We’re excited to release Muse Spark 1.3, which delivers improved performance across agentic and coding tasks. Drawing on what we learned from months of broad adoption of Muse Code and Meta Model API, we’ve also made this model easier to use in real-world settings. Smarter and more practically useful, Muse Spark 1.3 advances our work toward personal superintelligence.

Muse Spark 1.3 is rolling out today in Muse Code and Meta Model API. Previously available reasoning modes are available today with max reasoning coming shortly after we finish additional safety testing.

For more details about our evaluations, see our report.

Agentic Workflows

Muse Spark 1.3 is designed to better sustain longer-horizon work by collaborating with users and juggling multiple workflows in a single, long thread. When given an open-ended objective, it uses tools to generate its own context across messy and conflicting sources, proactively corrects gaps in its plan, and keeps track of what it has learned to produce a final deliverable. We trained the model across a diverse set of harnesses to generalize to various agentic environments.

Trained to more actively collaborate with the user, Muse Spark 1.3 asks clarifying questions when prompts are ambiguous, invokes help from the user when stuck, and confirms before taking consequential actions. When working on long tasks, it adapts to user preferences, either providing frequent updates or working silently in the background.

Muse Spark 1.3 follows complex, long-form instructions more reliably than earlier Muse Spark models. Across multi-step tasks, it’s better at preserving detailed requirements without dropping constraints or drifting from the requested workflow.

Note: this AI agent prototype was created by Muse Spark and is not a real product

We’ve also improved the multitasking capabilities of Muse Spark 1.3. For example, it now more accurately maps incoming prompts to the correct task within messy, single-threaded contexts, regardless of whether the user is steering past requests or interrupting them.

The model has better awareness of its own capabilities and limitations. We trained Muse Spark 1.3 to have a better sense of what it can and can’t do, what it knows and doesn’t know, and when it hits hurdles instead of hallucinating outcomes.

Availability

Muse Spark 1.3 is available today in Muse Code and in Meta Model API.

Get started with Muse Code

Safety

We’ve improved safety along several axes most relevant to agentic and coding capabilities. Muse Spark 1.3 shows stronger adversarial robustness, with improved resistance to adversarial inputs and prompt injections. On complex agentic tasks, the model has better calibration on what constitutes irreversible actions and proceeds accordingly. Together, these changes reflect better discretion and judgment in long-horizon agentic tasks.

Looking Forward

We have an exciting roadmap lined up, including bigger models, the Muse Spark open weights release, and more. Stay tuned.

联系我们 contact @ memedata.com