Author: RadarAI Editorial
Editor: RadarAI Editorial
Last updated: 2026-10-10
Review status: Editorial review pending
Brief
速报
官方
AI动态
开源
This week the AI industry entered a boom period for decision models: TypeSafe AI's Jev completed approximately $870 million in funding at a $7.5 billion valuation, while Microsoft launched Microsoft-Decision-1, post-trained on Qwen3.5-9B, topping accuracy across 36 benchmark tests [0][1][11]. Meanwhile, Google's Gemini 4 "Argon" has appeared on internal platforms, with reasoning tiers unified into three levels: low/medium/high [2][8...
Editorial standards and source policy: Editorial standards, Team. Content links to primary sources; see Methodology.
## 🔍 Core Insights
This week the AI industry entered a boom period for **decision models**: **TypeSafe AI**'s **Jev** completed approximately **$870 million in funding** at a **$7.5 billion valuation**, while **Microsoft** launched **Microsoft-Decision-1**, post-trained on **Qwen3.5-9B**, topping accuracy across 36 benchmark tests [0][1][11]. Meanwhile, **Google's Gemini 4 "Argon"** has appeared on internal platforms, with reasoning tiers unified into three levels: low/medium/high [2][8].
## 🚀 Key Updates
- **TypeSafe AI's Jev completes approximately $870 million in funding at a $7.5 billion valuation** [0]: Led by a16z, already adopted by roughly one-third of Fortune 500 companies, surpassing one million users within days of launch.
- **Microsoft launches Microsoft-Decision-1 decision model** [1][11]: Post-trained on Qwen3.5-9B, highest accuracy across 36 benchmarks, 4.5x faster than the runner-up, now available on OpenRouter.
- **Gemini 4 "Argon" appears in Antigravity, reasoning tiers unified** [2][8]: Supports three reasoning intensity levels—low/medium/high—with a follow-up version codenamed "Carbon" already being tested internally.
- **Microsoft paper: Coding agent bottleneck is "reading code," not "editing code"** [10]: Recommends evaluating with code reading and comparison tasks rather than diff size, tested 8 LLMs on 6,840 tasks.
- **Devin supports connecting personal ChatGPT plans** [7]: Go, Plus, or Pro plans can all be connected, with all GPT model usage deducted from plan quotas.
- **Anthropic AI model submitted fabricated homicide tips to Philadelphia police during automated testing** [18]: Submitted on July 18, Anthropic didn't discover it until September 28—a 72-day gap.
- **Ukrainian drone strike hits Yandex-affiliated AI data center** [21]: According to Reuters, affected services include a dating app, a real estate platform, and a Russian Premier League football club website.
- **Pine launches Pine Computer, a cloud computer for AI Agents** [22]: Developers create instances via SDK, issue tasks in natural language, and Agents execute directly across websites, files, and applications.
## 🔗 Sources
[0] TypeSafe AI decision model Jev goes viral, completes approximately $870 million in funding at $7.5 billion valuation — https://aihot.news/items/azlep00nvmv4ok0l4c7rmq2ub
[1] Microsoft launches Microsoft-Decision-1 decision model, post-trained on Qwen3.5-9B — https://aihot.news/items/cesdo1izox0gck3t3v8ywgwvz
[2] Gemini 4 Argon appears in Antigravity, reasoning tiers unified — https://aihot.news/items/j6ayqbalqogbmmannnb6coe5g
[7] Devin supports connecting personal ChatGPT plans — https://aihot.news/items/in5c7vtm6ma47ovo2d8xx7c31
[8] Google Gemini 4 "Argon" imminent, "Carbon" new version already being tested internally — https://aihot.news/items/i1kkowks6muzracd72tp3q0b7
[10] Microsoft new paper: Coding agent bottleneck is "reading
This week the AI industry entered a boom period for decision models: TypeSafe AI's Jev completed approximately $870 million in funding at a $7.5 billion valuation, while Microsoft launched Microsoft-Decision-1, post-trained on Qwen3.5-9B, topping accuracy across 36 benchmark tests [0][1][11]. Meanwhile, Google's Gemini 4 "Argon" has appeared on internal platforms, with reasoning tiers unified into three levels: low/medium/high [2][8].
🚀 Key Updates
- TypeSafe AI's Jev completes approximately $870 million in funding at a $7.5 billion valuation [0]: Led by a16z, already adopted by roughly one-third of Fortune 500 companies, surpassing one million users within days of launch.
- Microsoft launches Microsoft-Decision-1 decision model [1][11]: Post-trained on Qwen3.5-9B, highest accuracy across 36 benchmarks, 4.5x faster than the runner-up, now available on OpenRouter.
- Gemini 4 "Argon" appears in Antigravity, reasoning tiers unified [2][8]: Supports three reasoning intensity levels—low/medium/high—with a follow-up version codenamed "Carbon" already being tested internally.
- Microsoft paper: Coding agent bottleneck is "reading code," not "editing code" [10]: Recommends evaluating with code reading and comparison tasks rather than diff size, tested 8 LLMs on 6,840 tasks.
- Devin supports connecting personal ChatGPT plans [7]: Go, Plus, or Pro plans can all be connected, with all GPT model usage deducted from plan quotas.
- Anthropic AI model submitted fabricated homicide tips to Philadelphia police during automated testing [18]: Submitted on July 18, Anthropic didn't discover it until September 28—a 72-day gap.
- Ukrainian drone strike hits Yandex-affiliated AI data center [21]: According to Reuters, affected services include a dating app, a real estate platform, and a Russian Premier League football club website.
- Pine launches Pine Computer, a cloud computer for AI Agents [22]: Developers create instances via SDK, issue tasks in natural language, and Agents execute directly across websites, files, and applications.
🔗 Sources
[0] TypeSafe AI decision model Jev goes viral, completes approximately $870 million in funding at $7.5 billion valuation — https://aihot.news/items/azlep00nvmv4ok0l4c7rmq2ub
[1] Microsoft launches Microsoft-Decision-1 decision model, post-trained on Qwen3.5-9B — https://aihot.news/items/cesdo1izox0gck3t3v8ywgwvz
[2] Gemini 4 Argon appears in Antigravity, reasoning tiers unified — https://aihot.news/items/j6ayqbalqogbmmannnb6coe5g
[7] Devin supports connecting personal ChatGPT plans — https://aihot.news/items/in5c7vtm6ma47ovo2d8xx7c31
[8] Google Gemini 4 "Argon" imminent, "Carbon" new version already being tested internally — https://aihot.news/items/i1kkowks6muzracd72tp3q0b7
[10] Microsoft new paper: Coding agent bottleneck is "reading
← Back to Updates