Weekly review

Weekly review · Sep 28 – Oct 4

Signals published Sep 28 – Oct 4 (Monday to Sunday), by priority. Derived entirely from published signals: no editor's text, no AI.

Signal titles are in English where the source is; the editors' weekly summary is written in Chinese.

Signals this week10
Immediate attention0
Worth tracking10
Topics covered14

Worth tracking

10 signals

This week's changes worth keeping an eye on while they are confirmed.

2026-10-03· Hugging Face BlogWorth tracking

The Agent Said It Was Done. The Database Disagreed.

The Agent Said It Was Done. The Database Disagreed.

2026-10-02· Ollama ReleasesWorth tracking

Ollama v0.35.1: Clef decision models

Clef decision models Ollama now supports Clef and Clef Flash, Cloudflare's new open-source decision models, through /v1/systemone. Clef (27B) and Clef Flash (9B) are multimodal: requests can now include images alongside the text state, shared by all questions…

2026-10-01· InfoQ 中文Worth tracking

Gemini 4 Argon:DeepSWE 等多项领先,先只给网络安全团队用

Google 10 月 1 日发布 Gemini 4 Argon,在 DeepSWE(77.9%)、AutomationBench、法律与金融智能体评测上领先 GPT-6 Astra 与 Opus 5.5,输出上限从 6.4 万 token 提到 100 万。但它在 Terminal-bench 4.0 上只有 57.4%,低于 Opus 5.5 的 66.4% 和 Sonnet 5.5 的 70.6%,在 FrontierSWE 和 Terminal-Bench Science 上也落后于 Astra。首发价每百万 token 输入 2 美元、输出 10 美元,之后涨到 4 / 20 美元;眼下只通过 Fairwind 计划开放给网络安全防御团队,普通开发者还用不上。

2026-10-01· Hugging Face BlogWorth tracking

Introducing Olmo-core 3: Open, scalable training infrastructure for large MoEs

Introducing Olmo-core 3: Open, scalable training infrastructure for large MoEs

2026-09-30· InfoQ 中文Worth tracking

DeepSeek 开源昇腾版 DeepGEMM 等组件:接口与 NVIDIA 版相同

DeepSeek 9 月 30 日开源了一组跑在华为昇腾上的基础组件:算子编译工具 TileLang 的昇腾支持、矩阵乘法库 DeepGEMM-Ascend、专家并行通信库 DeepEP-Ascend,以及 TileKernels、FlashMLA、DeepSelect 等算子。DeepGEMM-Ascend 按 MIT 许可发布,与原版 DeepGEMM 接口完全兼容,在昇腾 950DT 上实测 BF16 达到 431 TFLOPS(硬件利用率 99.8%),FP4 达到 1701 TFLOPS。目前只在 950 系列上开发和验证,没有提到 910B / 910C。

2026-09-30· InfoQ 中文Worth tracking

Google 开源 AX:按 Kubernetes 的方式编排智能体,可秒级暂停

Google 9 月 30 日开源 AX,一个按 Kubernetes 方式声明式编排自主智能体的系统,Apache 2.0 许可。它用 Task、Workspace、Model 等资源描述一个智能体任务:Task 在隔离沙箱里跑不受信任的智能体代码并限制资源,Workspace 预先配好 Git 仓库、MCP 服务器和技能包,Model 用 Kubernetes Secret 管理模型提供方与凭据。任务可以在亚秒级暂停和恢复,多个任务复用同一组工作进程。仓库说明写明仍在重度开发中,稳定版之前会有大的不兼容改动。

2026-09-29· OpenAI NewsWorth tracking

Introducing GPT-6.1 Sol

Meet GPT-6.1 Sol: near-Astra intelligence for coding, computer use, and professional work at one-fifth of Astra’s standard API input and output token prices.

2026-09-29· Hugging Face BlogWorth tracking

NVIDIA Kumo Tabular Sets a New Accuracy-Efficiency Frontier for Tabular Prediction

NVIDIA Kumo Tabular Sets a New Accuracy-Efficiency Frontier for Tabular Prediction

2026-09-28· Hugging Face BlogWorth tracking

Holo4: powering generalist computer-use agents

H company released Holo4, a family of open-weight agentic models for computer use across GUIs, code, MCP and APIs: Holo4-27B (dense), Holo4-35B-A3B (MoE) and Holotron4 Nano (30B-A3B). Holo4-27B scores 61.7% on OSWorld 2.0.

2026-09-28· InfoQ 中文Worth tracking

Claude Sonnet 5.5:终端任务从 10% 到 71%,按任务算未必更便宜

Anthropic 9 月 28 日发布 Claude Sonnet 5.5,每百万 token 输入 2 美元、输出 10 美元,是 Opus 5.5 的一半,也与 GPT-6 Sol 同价。它在 Terminal-Bench 4.0 上从 Sonnet 5 的 10.3% 升到 70.6%,超过 Opus 5.5 的 66.4%,但在 FrontierCode 与 OSWorld 上仍略低于 Opus 5.5。Anthropic 称单任务成本最多低 30%;而 Artificial Analysis 在最高推理档测得单任务 7.60 美元,比 Sonnet 5 高约 50%。

Past weeks

16 weeks

Subscribe

One item a week