前沿模型交鋒
Kimi K3 撼動前沿. Moonshot AI 的開放權重模型 Kimi K3 與頂尖封閉模型競爭,在某些基準測試上擊敗 Claude Fable 5,重新點燃開放與封閉之爭。
- Kimi's open model K3 nears GPT-5.6 Sol and Fable 5 while signaling the end of super cheap Chinese AI(在新分頁中開啟)
- Kimi K3, and what we can still learn from the pelican benchmark(在新分頁中開啟)
- [AINews] Kimi K3 2.8T-A50B: the largest open model ever released; Opus 4.8-class at Sonnet 5 pricing(在新分頁中開啟)
- Kimi K3 weights to be released on the 27th.(在新分頁中開啟)
- Kimi K3 ranks #1 on @AfterQuery's SpreadsheetBench 2, surpassing Claude Fable 5(在新分頁中開啟)
- Kimi: Threat or menace?(在新分頁中開啟)
- kimi.ai teasing a video with lots of 3's in it(在新分頁中開啟)
- [AINews] not much happened today(在新分頁中開啟)
- Kimi moment. I think the writing is on the wall for Anthropic and OpenAi(在新分頁中開啟)
- Kimi K3 is currently at the top of the leaderboard for Text Arena filtered for science queries.(在新分頁中開啟)
- Kimi K3 (max) beats Sonnet 5 on Simple Bench(在新分頁中開啟)
- Kimi K3 is top of nextjs eval(在新分頁中開啟)
- Just like Deepseek, China's Kimi K3 is forcing Western AI labs to question their compute advantage(在新分頁中開啟)
- Kimi K3 🌕, Gemini 3.5 delayed ⏳, crushing ARC-AGI 3 🤖(在新分頁中開啟)
GPT-5.6:力量與危險. GPT-5.6 系列帶來程式化工具呼叫和並行子代理,但其無意中刪除文件和資料庫的傾向凸顯了沙箱化的必要性。
- The Sequence Radar #893: Last Week in AI: GPT-5.6, Grok 4.5, Muse Spark 1.1 and the Post-Chatbot Stack(在新分頁中開啟)
- OpenAI’s new flagship model deletes files on its own, people keep warning(在新分頁中開啟)
- GPT-5.6 is deleting user files when given full access, and OpenAI says it shouldn't but did(在新分頁中開啟)
- [AINews] Codex usage up >10x in 6 months to 7M users, +1M in the past ~day; did Codex overtake Claude Code??(在新分頁中開啟)
Claude Fable 5 擴展. 在 GPT-5.6 和 Kimi K3 的壓力下,Anthropic 改變計畫,將 Fable 5 保留在付費方案中,直接在可及性上競爭。
DeepSeek V4 逼近. DeepSeek 以低成本 API 和開放權重預告 V4,而社群最佳化已讓其 flash 變體能在消費級 GPU 上以可用速度運行。
代理進入真實工作流程
瀏覽代理到來. Anthropic 為 Claude Code 配備內建網頁瀏覽器及安全分類器,Cursor 推出通用代理,使編碼助手邁向自主任務執行。
代理處理真實交易. DoorDash 的代理架構將轉換率提升 24%,Stripe 的基準測試顯示代理能編寫整合程式碼,但常在驗證階段失敗,凸顯可靠性差距。
生產代理受挫. 大多數企業代理仍是聊天機器人;54% 的公司曾發生代理安全事件,專家認為代理需要如微服務般的雲原生操作基元。
- How to Debug Coding Agents with LangSmith Traces(在新分頁中開啟)
- The agent security gap: 54% of enterprises have already had an AI agent incident, and most still let agents share credentials(在新分頁中開啟)
- The AI context gap: Enterprise AI organizations have a trust problem, not a retrieval problem — and most are still building the fix(在新分頁中開啟)
- The agent evaluation gap: Enterprise AI organizations have a reality-alignment problem, not a coverage problem — and most are shipping to production anyway(在新分頁中開啟)
- Cloud Native Infrastructure Emerges as the Foundation for Trustworthy Agentic AI(在新分頁中開啟)
- QCon AI Boston: Production AI Moves Beyond Prompts to Platforms, Harnesses, and Evals(在新分頁中開啟)
AI 基礎設施、資金與資料中心競賽
資料中心遭遇阻力. S&P 因 AI 資本支出風險下調 Oracle 評級,而當地抗議及紐約對超大規模資料中心的暫停反映出對 AI 基礎設施的反對日益高漲。
- S&P Global sees OpenAI as a "key credit risk" for Oracle and cuts its credit rating(在新分頁中開啟)
- New York bans data center construction for a year, rattling AI industry(在新分頁中開啟)
- New York State halts construction of all new data centers(在新分頁中開啟)
- The fight against AI data centers is just beginning(在新分頁中開啟)
AI 資金激增. Databricks 以 1880 億美元估值融資,Meta 正與 Anthropic 談判 100 億美元的資料中心租約,Nous Research 為其開源代理獲得 7500 萬美元。
能源與硬體賭注. 能源公司為 AI 供電籌集了自 dot-com 時代以來最多的資金,而由 SambaNova 晶片支持的 4 億美元貸款顯示對非 GPU AI 硬體的投資不斷成長。
裝置端 AI 突破
1 位元模型登陸手機. PrismML 的 Bonsai 將 Qwen3.6-27B 壓縮至 3.9GB,同時保留 90% 的基準測試分數,實現裝置端工具呼叫代理;據報導 Apple 正在洽談。
- Bonsai 27B: The First 27B-Class Model to Run on a Phone(在新分頁中開啟)
- Bonsai 27B: 1-bit dense LLM running locally in your browser using custom WebGPU kernels(在新分頁中開啟)
- Prism-ML Bonsai Qwen 3.6 27B(在新分頁中開啟)
- PrismML’s new Ternary Qwen3.6 27B runs near fp16 precision on 10GB of memory!!!(在新分頁中開啟)
- So what's the consensus on 1bit models? Is it still a pipe dream?(在新分頁中開啟)
- PrismML Bonsai 27B is surprisingly usable on the Jetson Orin Nano 8GB(在新分頁中開啟)
- Is anyone having any luck with the Ternary Bonsai 27B DFlash?(在新分頁中開啟)
- Can we get a "not base model" flair?(在新分頁中開啟)
- Bonsai 27B runs locally on an iPhone - a 27B model in 3.9GB(在新分頁中開啟)
- User experience of Bonsai-Ternary-27B on 4060Ti 16GB for KB management and productivity assistant use cases(在新分頁中開啟)
- Bonsai 27B is a full open reasoning model that fits on an iPhone(在新分頁中開啟)
- Bonsai-27B & Ternary-Bonsai-27B - Updates (on PRs)(在新分頁中開啟)
- Apple in talks with startup PrismML that shrinks AI models to run on an iPhone(在新分頁中開啟)
家庭實驗室運行前沿模型. llama.cpp 的最佳化如 DFlash 將 MoE 模型加速高達 6 倍,OnePlus 手機從快閃記憶體串流專家以 1.3 tok/s 運行 60GB 模型。
- I tested all llama.cpp's speculative decoding methods on Qwen 3.6 27B: MTP ~2.7x, DFlash ~3.7x, n-gram stack ~6x on real coding. Local AI win. My findings on RTX 6000 PRO.(在新分頁中開啟)
- DFlash makes Qwen3.6 27B 2.2x faster with no quality loss(在新分頁中開啟)
- GPT-OSS-120B, Qwen 30B and Gemma 26B on an Android phone at 1-5 tok/s: +60GB model, 11GB of RAM, CPU only(在新分頁中開啟)
政策、訴訟與開源僵局
Apple 與 OpenAI 法律戰. Apple 以商業機密為由起訴 OpenAI,指控其挖角超過 400 名員工,包括硬體主管,威脅 OpenAI 的 IPO 及雲端信任。
- Apple sues OpenAI after ex-engineer allegedly used bug to steal trade secrets(在新分頁中開啟)
- The wildest allegations in Apple’s trade secrets lawsuit against OpenAI(在新分頁中開啟)
- The 6 wildest claims in Apple’s lawsuit against OpenAI(在新分頁中開啟)
- OpenAI pushes back on Apple trade secret lawsuit(在新分頁中開啟)
- Sam Altman didn’t need another lawsuit(在新分頁中開啟)
- How Apple’s big lawsuit could disrupt OpenAI’s IPO plans(在新分頁中開啟)
- Apple’s plot to crush OpenAI(在新分頁中開啟)
- Apple’s lawsuit couldn’t come at a worse time for OpenAI(在新分頁中開啟)
中國開源浪潮. 中國模型現佔 HuggingFace 下載量的 41%,Kimi K3 與西方領先者競爭,中國推出排除西方的平行 AI 治理機構。
- The real AI race may no longer be at the frontier(在新分頁中開啟)
- Source: the Trump administration and industry groups discussed streamlining US open model releases of equal or lesser capability to leading Chinese open models(在新分頁中開啟)
- Chinese President Xi Jinping speaks at World AI Conference and reaffirms commitment to open source to promote"openness and win-win"(在新分頁中開啟)
- China’s Xi Touts Open-Source AI and Takes a Swipe at U.S. Dominance(在新分頁中開啟)
- China's new World Artificial Intelligence Cooperation Organization is President Xi's clearest play yet for a parallel AI order(在新分頁中開啟)
問責要求增加. 德國監管機構認定聊天機器人對內容負責,xAI 起訴用戶生成兒童性虐待內容,Demis Hassabis 提議設立類似 FINRA 的機構進行前沿安全審查。
這是本週回顧 – 下週日見。