PulseAugur
实时 22:02:25
English(EN) Kimi K3 API: 2.8T Parameters at 1/3 the Cost of GPT-5

Moonshot AI 发布 Kimi K3,拥有 2.8T 参数,成本挑战 GPT-5.6

Moonshot AI 推出了其 Kimi K3 模型,这是一个拥有 2.8 万亿参数的混合专家(Mixture-of-Experts)架构,具有 100 万 token 的上下文窗口。该模型被定位为 GPT-5.6Claude Fable 5 等竞争对手的经济高效替代品,API 使用成本显著降低,尤其适用于高流量应用。Kimi K3 还拥有新颖的提示缓存机制,可进一步降低重复上下文的成本,并且可以通过 Yingsuan AI 统一网关访问,该网关与 OpenAI SDK 兼容。 AI

影响 为大上下文窗口模型树立了新的价格-性能基准,可能降低 AI 应用的成本。

排序理由 前沿实验室模型发布,附带系统卡和基准数据。[lever_c_demoted from frontier_release: ic=1 ai=1.0]

在 dev.to — LLM tag 阅读 →

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

Moonshot AI 发布 Kimi K3,拥有 2.8T 参数,成本挑战 GPT-5.6

本文如何被排名

Signal score
3 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Significant
前沿实验室模型发布,附带系统卡和基准数据。[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
Same-day
Cluster formed today. Ranking reflects the current source set at time of score.

完整方法见我们的编辑标准

报道来源 [1]

  1. dev.to — LLM tag TIER_1 English(EN) · YingSuan AI ·

    Kimi K3 API:成本仅为GPT-5的三分之一,拥有2.8T参数

    <h1> Kimi K3 API: 2.8T Parameters at 1/3 the Cost of GPT-5 </h1> <p>Let’s cut through the hype: Moonshot AI’s <strong>Kimi K3</strong>, launched in July 2026, isn’t just another frontier model. It’s a 2.8-trillion-parameter Mixture-of-Experts (MoE) architecture with a 1M-token co…