2026-10-04 AI 日报

开源项目

1. Niko1221/Strata

   Qwen3.8-Flash-Next on any consumer hardware: one-click install for Windows / Linux. Strata inference engine, OpenAI/Anthropic API on localhost, optional image…

2. earthtojake/text-to-cad

   Give your agent CAD superpowers.

3. calesthio/OpenMontage

   World's first open-source, agentic video production system. 12 production pipelines, 100+ tools, 700+ agent skill and production-knowledge files. Turn your AI…

4. Panniantong/Agent-Reach

   Give your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.

5. DietrichGebert/ponytail

   Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.

6. coreyhaines31/marketingskills

   Marketing skills for Claude Code and AI agents. CRO, copywriting, SEO, analytics, and growth engineering.

7. Nasiko-Labs/nasiko

   The Open Runtime for AI Agents

8. thedotmack/claude-mem

   Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context…

9. mostlygeek/llama-swap

   Reliable model swapping for any local OpenAI/Anthropic compatible server - llama.cpp, vllm, etc

10. ys-ll/uniterm

    A lightweight all-in-one terminal with 30+ protocols — SSH, RDP, SFTP, databases, Kubernetes and more. With a built-in autonomous AI Agent that plans and runs…

论文

本统计日暂无条目。

行业新闻

1. Run Qwen 3.8 Flash Next (125B) on consumer hardware (RTX 4090) at 100T/s

2. "Torturing" LLMs in a Robot Prison Has Triggered the Dumbest Debate in AI Yet

社区动态

1. Top ARC-ΑGI-3 scores on Kaggle just went from 7% to 56% [N]

   https://preview.redd.it/gkjgii48gfth1.png?width=575&format=png&auto=webp&s=4e1c35bdacf41d18a6409baedcddc9432ec79814 This happened over the past 30 days. So…

2. An LLM scoreboard for whether an AI falls for the same trick questions people do

   https://preview.redd.it/t0osnuw50fth1.png?width=1401&format=png&auto=webp&s=a0924bdfdccf1c81f096af95c9200b33410074d2 People are bad at a few famous trick…

3. Working with an AI Company That Does Things You Disagree With [D]

   I'm a PhD student in machine learning in the EU and was looking for internships at exciting companies. I shortlisted few and applied by reaching out to people…

4. Local text to speech with Breeze is truly incredible

   Been playing around with local TTS with Breeze combined with STT, and the results are amazing. Using Opus 5.5, I can hear the first sound after 500ms if there…

5. Day 4 no Gemini 4

   Daily reminder we are paying monthly for AIs with knowledge cut offs from almost two years ago.

6. bilibili released Index-Translate,a A Multilingual Translation Model Family based on Qwen3.5

   https://github.com/bilibili/Index-Translate Index-Translate is a family of multilingual translation models built on Qwen3.5. The text models cover 150…

7. Agentic behavior

   Since around a year ago, my business built our own agent software running on our FreeBSD servers. We selected to use the Anthropic models through their API…

8. the official ICLR template .bib has had Bengio listed twice since 2019 [D]

   i work on reference checking stuff so i was reading through the ICLR 2027 author guidelines and style files this week the sample .bib that ships with the…

9. Need maybe say "Use llama.cpp"

   So I tried that miracle engine everyone is talking about. Asked the IQ3_S model to express its opinion on a post from this sub to measure the tps on a…

10. Everyone is obsessed with trillion-parameter models, so I mapped out the entire AI spectrum from 100KB to 2.5TB (and what they actually cost to run)

    Right now, the AI space feels entirely focused on massive datacenter clusters and renting H100s by the hour. But after spending way too much time looking at…

官方发布

1. 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.

2. Gemini 4 Argon: our next era of frontier intelligence

   Announcing Gemini 4 Argon, our frontier model for real-world coding, enterprise knowledge work, and cyber defense, rolling out soon.

3. We've launched Claude Sonnet 5.5 (claude-sonnet-5-5).

   We've launched Claude Sonnet 5.5 (claude-sonnet-5-5). It's available on the Claude API, Claude in Amazon Bedrock, Claude Platform on AWS, Claude on Google…

4. Basis completes a tax workbook 2x faster with GPT-6 Astra

   GPT-6 Astra completed a 50-tab tax workbook twice as fast as GPT-5.6 Sol, and its stronger understanding of user intent gives Basis more confidence in…

5. Introducing dots

   Dots by OpenAI are proactive assistants that can keep working across complex projects and everyday tasks. Learn how dots help you stay in control while work…

6. Z.ai GLM 5.3 (zai-glm-5-3) is now Generally Available.

   Z.ai GLM 5.3 (zai-glm-5-3) is now Generally Available.

7. Claude-shaped science

   Guest author Prof. Matthew Schwartz describes what happened when he stopped fighting Claude and allowed Claude to find “Claude-shaped” problems: ones best…

8. grok-voice-transcribe-1.0 end of life

   grok-voice-transcribe-1.0 is deprecated and reaches end of life on October 2, 2026. All requests to that slug are routed to grok-voice-transcribe-2.0 at the…

9. GLM-5.3 and the spread of advanced cyber capabilities

   Like Claude Mythos Preview, GLM-5.3 has strong capabilities for autonomously building end-to-end cyber exploits. But GLM-5.3 is unlike other frontier models…

10. Let skills in Gemini tackle your most repetitive tasks

    Now you can automate your most repetitive tasks more easily with reusable custom instructions through skills, which will be replacing gems.