Author: RadarAI Editorial
Editor: RadarAI Editorial
Last updated: 2026-10-24
Review status: Editorial review pending
Brief
速报
官方
AI动态
开源
Moonshot AI releases Kimi K3.1 Pro, challenging the ceiling of open-weight models with a 2.8T total parameter / 104B activated MoE architecture and 1M context, ranking near the top on AA-AgentPerf and the cybersecurity index [7][11][12][14]; meanwhile, the AI layoff wave has spread from American Express and Barclays to eBay and Netflix, as AI's displacement effect on white-collar jobs turns from prediction into financial-report-level reality [10].
Editorial standards and source policy: Editorial standards, Team. Content links to primary sources; see Methodology.
## 🔍 Core Insights
**Moonshot AI** has released **Kimi K3.1 Pro**, challenging the ceiling of open-weight models with a **2.8T total parameter / 104B activated** MoE architecture and **1M context**, ranking near the top on **AA-AgentPerf** and the cybersecurity index [7][11][12][14]; meanwhile, the **AI layoff wave** has spread from American Express and Barclays to eBay and Netflix, as **AI's displacement effect on white-collar jobs** turns from prediction into financial-report-level reality [10].
## 🚀 Key Updates
- **Kimi K3.1 Pro officially released** [7][11][14]: 3T-class MoE, 1M context, with a simultaneous 2–3x speed Highspeed version
- **Kimi K3.1 Pro tops the open-model cybersecurity index** [12]: scoring 64%, second only to GPT-6 Sol's 69%
- **Artificial Analysis evaluates Kimi K3.1 Pro** [6]: leads in knowledge work and coding, AA-3D preference voting still to be observed
- **Meta Muse Spark Open inference throughput benchmark goes live** [3]: 816B total parameters, 42B activated, integrated into AA-AgentPerf
- **Jack Dorsey releases full-duplex voice model Dessa** [8]: trained purely on voice data, with emotional expression and naturalness surpassing existing models
- **OpenAI discloses Brotli unauthenticated RCE vulnerability** [9]: affects embedded components in Node.js, Bun, Envoy, and others, discovered during a red-team exercise
- **Four major companies announce AI-related layoffs in quick succession** [10]: American Express cuts 3,500, Barclays cuts 5,000, eBay cuts 9%, Netflix cuts 6%
- **Claude Code v2.1.310 released** [2]: adds Monitor tool and plugin monitoring, covering Bedrock, Vertex, and Foundry
- **Nvidia CEO boosts Microduck robot** [20]: pre-orders exceed 20,000 units after GTC Berlin demo, delivery before Christmas
## 🔗 Sources
[1] Indigo Talk recording: Wan Weigang and Lao Yu discuss the AI era — https://aihot.news/items/j9qeassxv2qm6tng5wvhu2h0b
[2] Claude Code v2.1.310 released: adds Monitor tool and plugin monitoring — https://aihot.news/items/uckmhjx55injing6ui7hr7j1i
[3] Artificial Analysis publishes Meta Muse Spark Open inference throughput benchmark results on AA-AgentPerf — https://aihot.news/items/fmiy7354owk8sx92qygn846jp
[4] fal.ai website fully redesigned and relaunched — https://aihot.news/items/t917qbffpnmtwlh0kjphdt5l6
[5] Several minor Codex updates: faster call connections — https://aihot.news/items/qu5md7zglhhizmsz2fapm646n
[6] Artificial Analysis evaluates Kimi K3.1 Pro: leads in knowledge work and coding, AA-3D may rank lower — https://aihot.news/items/k3zzutim37x5zypu0virne136
[7] Moonshot AI releases Kimi K3.1 Pro: 3-trillion-class parameters, supports 1M context — https://aihot.news/items/q7fw
Moonshot AI has released Kimi K3.1 Pro, challenging the ceiling of open-weight models with a 2.8T total parameter / 104B activated MoE architecture and 1M context, ranking near the top on AA-AgentPerf and the cybersecurity index [7][11][12][14]; meanwhile, the AI layoff wave has spread from American Express and Barclays to eBay and Netflix, as AI's displacement effect on white-collar jobs turns from prediction into financial-report-level reality [10].
🚀 Key Updates
- Kimi K3.1 Pro officially released [7][11][14]: 3T-class MoE, 1M context, with a simultaneous 2–3x speed Highspeed version
- Kimi K3.1 Pro tops the open-model cybersecurity index [12]: scoring 64%, second only to GPT-6 Sol's 69%
- Artificial Analysis evaluates Kimi K3.1 Pro [6]: leads in knowledge work and coding, AA-3D preference voting still to be observed
- Meta Muse Spark Open inference throughput benchmark goes live [3]: 816B total parameters, 42B activated, integrated into AA-AgentPerf
- Jack Dorsey releases full-duplex voice model Dessa [8]: trained purely on voice data, with emotional expression and naturalness surpassing existing models
- OpenAI discloses Brotli unauthenticated RCE vulnerability [9]: affects embedded components in Node.js, Bun, Envoy, and others, discovered during a red-team exercise
- Four major companies announce AI-related layoffs in quick succession [10]: American Express cuts 3,500, Barclays cuts 5,000, eBay cuts 9%, Netflix cuts 6%
- Claude Code v2.1.310 released [2]: adds Monitor tool and plugin monitoring, covering Bedrock, Vertex, and Foundry
- Nvidia CEO boosts Microduck robot [20]: pre-orders exceed 20,000 units after GTC Berlin demo, delivery before Christmas
🔗 Sources
[1] Indigo Talk recording: Wan Weigang and Lao Yu discuss the AI era — https://aihot.news/items/j9qeassxv2qm6tng5wvhu2h0b
[2] Claude Code v2.1.310 released: adds Monitor tool and plugin monitoring — https://aihot.news/items/uckmhjx55injing6ui7hr7j1i
[3] Artificial Analysis publishes Meta Muse Spark Open inference throughput benchmark results on AA-AgentPerf — https://aihot.news/items/fmiy7354owk8sx92qygn846jp
[4] fal.ai website fully redesigned and relaunched — https://aihot.news/items/t917qbffpnmtwlh0kjphdt5l6
[5] Several minor Codex updates: faster call connections — https://aihot.news/items/qu5md7zglhhizmsz2fapm646n
[6] Artificial Analysis evaluates Kimi K3.1 Pro: leads in knowledge work and coding, AA-3D may rank lower — https://aihot.news/items/k3zzutim37x5zypu0virne136
[7] Moonshot AI releases Kimi K3.1 Pro: 3-trillion-class parameters, supports 1M context — https://aihot.news/items/q7fw
← Back to Updates