PulseAugur
实时 01:42:08
English(EN) Cognition Releases SWE-2: A Kimi K3 Post-Trained Coding Model That Matches Fable 5.1 on FrontierCode at 64% Lower Cost

Cognition 的 SWE-2 编码模型以更低的成本匹配 Fable 5.1

Cognition 发布了 SWE-2,这是一款使用 Moonshot AIKimi K3 模型进行强化学习后训练的新型编码模型。据报道,该新模型在 FrontierCode 基准测试上的表现与 Fable 5.1 相当,但成本显著降低。SWE-2 还引入了可选的推理工作量级别,在一次强化学习运行中进行训练,并且与前代 SWE-1.7 相比,在编码任务上展现出更高的效率和更少的绕路。 AI

影响 为编码模型设定了新的成本效益基准,可能影响未来的开发和部署策略。

排序理由 来自前沿实验室 (Cognition) 的新模型发布,并附有性能基准。 [lever_c_demoted from frontier_release: ic=1 ai=1.0]

在 MarkTechPost 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →

Cognition 的 SWE-2 编码模型以更低的成本匹配 Fable 5.1

本文如何被排名

Signal score
53 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Significant
来自前沿实验室 (Cognition) 的新模型发布,并附有性能基准。 [lever_c_demoted from frontier_release: ic=1 ai=1.0]
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
model release, product
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准

报道来源 [1]

  1. MarkTechPost TIER_1 English(EN) · Asif Razzaq ·

    Cognition 发布 SWE-2:一款 Kimi K3 经后训练的编码模型,在 FrontierCode 上以 64% 的成本匹配 Fable 5.1

    <p>Cognition, the company behind the Devin coding agent, has released SWE-2, its most capable coding model to date. SWE-2 is post-trained with reinforcement learning from Kimi K3, Moonshot AI&#8217;s 2.8T-parameter open model. Cognition reports a score of 50.0% on FrontierCode 1.…