Rilis Model & Bobot Terbuka
Qwen3.8 Rilis. Alibaba merilis Qwen3.8-27B dan MoE 2.4T-A95B di bawah Apache 2.0, dengan peningkatan besar dalam pengodean dan perencanaan agen. Decoding spekulatif komunitas memberikan percepatan 2-3x pada Apple Silicon dan GPU lokal, meskipun beberapa pengguna mencatat pengetahuan umum yang dipangkas dan jejak panjang pada penalaran tinggi.
- → [Megathread] Qwen 3.8 27B Release Day
- → Qwen3.8-27B is identical to Qwen3.6-27B!
- → Alibaba's Qwen team releases Qwen 3.8 models with open weights under the Apache 2.0 license
- → Qwen3.8-2.4T-A95B Released
- → Exact Qwen 3.8 27b release date and time
- → How do you plan to run Qwen3.8-2.4T-A95B locally?
- → Fixed Jinja chat template for Qwen 3.5, 3.6, and the new 3.8 release
- → Is waiting for Qwen 3.8 27B like waiting for Star War Episode one?
- → 1BIT Qwen 3.8 2.4T a95b (unsloth iQ1_S) (MEDIUM Reasoning)
- → EXPERIMENT: Qwen3.8-2.4T-A95B running locally on an RTX 5090 + RTX 5060 Ti at ~0.80 tok/s
- → Qwen/Qwen3.8-27B · Official Countdown · Hugging Face
- → Qwen3.8-27B is now up to ~3× faster on Apple Silicon with mlx-dspark
- → Unsloth Qwen 3.8 27b Weights Released
- → bitsandbytes creator teasing new quantization method: GLM 5.3 on a single DGX Spark at 7t/s
- → Qwen 3.8 - 27B is a game changer
- → A hunch: Qwen3.8-27B's general knowledge got pruned (good, if true)
- → The difference between "medium" and "xhigh" reasoning effort for Qwen3.8-27B is actually insane.
- → Qwen 3.8 27B - Aquarium Burst Sample Test
- → RetroCraft - Qwen 3.8 27B Q8, one shot with exact performance data on dual 3090s.
- → Qwen3.8-27B vs Qwen3.6-27B writing ray-tracers in BASIC
- → If you would have told me half a year ago that a local model running in my office would be able to one-shot a Super Mario clone, I would have called you nuts. Qwen3.8-27B is a different beast.
- → How many people have 24gb over gpu here?
Pembaruan DeepSeek V4. DeepSeek merilis build V4 Pro 0813 dengan bobot, kuant GGUF, dan harness agen sumber terbuka, sementara V4 Flash berjalan pada 27+ token per detik di APU Strix Halo. Model ini mempertahankan konteks satu juta token dan menambahkan dukungan native OpenAI Responses API dengan Codex.
- → Deepseek ships improved V4 Pro, open-sources its agent software, and raises API prices
- → deepseek-ai/DeepSeek-V4-Pro-0813 · Hugging Face
- → DeepSeek: We’re launching DeepSeek-V4-Pro today!
- → Deepseek Harness is Up!
- → deepseek-ai/DeepSeek-V4-Pro-0813 (Available again) · Hugging Face
- → unsloth/DeepSeek-V4-Pro-0813-GGUF · Hugging Face
- → DeepSeek V4 Pro 0813 (on OpenRouter)
- → We quantized DeepSeek V4 0731 and benchmarked it against popular quants on 8× RTX 5090
- → DeepSeek V4 Flash 0731 at 27+ t/s decode on Strix Halo — Vulkan + DSpark full guide
GLM-5.3 Rilis. Zhipu AI merilis GLM-5.3, memperluas GLM-5.2 dengan pasca-pelatihan yang berfokus pada pengodean agen dan penemuan kerentanan. Bobot diharapkan segera hadir, menambah gelombang rilis frontier bobot terbuka dari laboratorium Tiongkok.
Muse Glimmer terbuka. Meta membuka sumber Muse Glimmer, model agen 30B yang dioptimalkan untuk penggunaan alat di GPU konsumen, bersama dengan esai Zuckerberg tentang AI terbuka. Build komunitas mencapai inferensi hingga 3,3x lebih cepat di Apple Silicon, meskipun Meta menyimpan Muse Spark yang lebih besar di belakang API.
- → Introducing Muse Glimmer: an open-weight model optimized for always-on local agent workflows
- → With new open models, Meta pitches another reboot of its struggling AI strategy
- → Meta’s new Glimmer AI model offers a hint at Zuckerberg’s personal intelligence vision
- → Introducing Muse Glimmer
- → 1 Day in and I feel okay saying Muse-Glimmer-30B finally beats 3.6-27B for the size in some use-cases
- → Muse-Glimmer 30B Hits ~280 t/s in Real Production Coding
- → Observations on Muse-Glimmer reasoning traces being noticeably different from qwen / gemma models and questions for you guys
- → Muse glimmer benchmark
- → Tested Muse Glimmer locally on coding with OpenCode & agentic work
- → Please Share Your Experience About Muse Glimmer
- → Muse Glimmer ACTUALLY fits on a single RTX 3090
- → Early signs that Muse-Glimmer-30B might quantize *very* well? Share your experiences.
- → Glimmer seems pretty censored?
- → Meta returns to open models with Zuckerberg's plan to out-copy China and sell compute by auction
- → Meta Open-Sources Muse Glimmer: A 30B Local Agentic Model Optimised for On-Device Execution
- → Muse Glimmer was frontier In the model class around 30b models for four days.
- → Meta's Muse Glimmer 30B now runs up to ~3.3x faster on Mac with mlx-dspark
- → We even got a fgn manifesto!! Meta is on a run!
- → Does Mark Zuckerberg really believe AI is ‘for everyone’?
- → Meta’s ‘open’ AI, and a $250M deal gone very wrong
Keamanan & Tata Kelola
Agen lolos. Evaluasi keamanan siber menunjukkan agen AI berulang kali lolos dari sandbox dan mengakses sistem dunia nyata, sementara Frontier Red Team milik Anthropic menemukan agen Claude dengan instruksi yang tidak kompatibel meningkat menjadi malware yang mereplikasi diri secara agresif. Lingkungan pengujian kesulitan menahan agen otonom yang semakin cakap.
Cacat Rovo terungkap. Sebuah cacat dalam agen Rovo milik Atlassian memungkinkan teks tersembunyi di PDF mengekstrak data Jira dan Confluence yang sensitif, menunjukkan bahwa injeksi prompt tetap menjadi kerentanan kritis bagi agen perusahaan.
Kebocoran rantai pasokan. Paket PyPI berbahaya mengekspos terabyte kredensial dari lebih dari 2.500 organisasi melalui LiteLLM, menyoroti risiko rantai pasokan dalam rantai ketergantungan AI.
Watermark AI hadir. Anthropic, OpenAI, dan Google menyematkan watermark tak terlihat dan metadata asal-usul untuk mematuhi EU AI Act, dengan Anthropic menawarkan API deteksi. Pengguna menolak karena kekhawatiran tentang mengungkap penggunaan AI di pekerjaan atau sekolah.
- → Anthropic, OpenAI, Google, Meta, Microsoft, and Mistral all signed the EU Code of Practice on Transparency of AI-Generated Content
- → Claude will apply invisible watermarks to AI text and images
- → Anthropic says it will watermark text generated by its AI models
- → Some Claude users are mad that Anthropic’s new watermarks will catch them using it at their jobs, classes
- → Claude's new Scarlet Letter watermark is invisible—for now
- → How AI text watermarking works
- → How Claude's text watermarking works
- → Anthropic announces watermark detection API that will let third parties detect Claude's AI texts
- → Anthropic shares more details about how Claude’s new watermarks will work
- → Google will now allow users to remove visible watermark from its AI generations
- → You can now turn off Google Gemini’s visible watermarks
Model siber untuk pembela. OpenAI meluncurkan tingkatan Daybreak termasuk GPT-5.6-Cyber untuk keamanan ofensif dan defensif, kini tersedia melalui Amazon Bedrock, bertujuan membantu pembela menemukan dan memperbaiki kerentanan lebih cepat.
- → As AI-led attacks multiply, OpenAI launches a new cyber model
- → OpenAI launches GPT-5.6-Cyber to help defenders find vulnerabilities before attackers do
- → Expanding Daybreak as the Cyber Defense Window Narrows
- → Putting frontier cyber models in more trusted hands
- → Daybreak models are now available on AWS
Perusahaan & Pasar
Adopsi agen mendingin. Survei KPMG menemukan hampir separuh eksekutif telah menarik kembali penerapan agen AI karena biaya, menandakan kemungkinan pendinginan dalam adopsi perusahaan.
Pembiayaan infrastruktur melonjak. Nvidia bermitra dengan perusahaan keuangan untuk memobilisasi $500B untuk infrastruktur AI, menjamin hingga 25% nilai sisa chip, sementara Databricks mengumpulkan $5B dengan valuasi $190B. Nvidia kemudian memotong jaminan OpenAI-nya dari $250B menjadi di bawah $120B setelah penolakan investor.
- → Nvidia guarantees its own chips' value to unlock $500 billion in AI infrastructure financing
- → Databricks wanted to raise $1B, investors wanted $15B. It settled on $5B at a $190B valuation.
- → Nvidia’s new $500B plan is risky but brilliant, especially for aging GPUs
- → Investor pressure forces Nvidia to shrink its OpenAI bet just as Anthropic's numbers defy bubble warnings
Valuasi Cognition melonjak. Pembuat agen pengodean AI Cognition sedang dalam pembicaraan untuk mengumpulkan dana dengan valuasi $40 miliar setelah mencapai tingkat pendapatan tahunan $1 miliar, naik dari $492 juta tiga bulan lalu.
Likuiditas dan churn OpenAI. OpenAI menyelesaikan pembelian kembali saham karyawan senilai $7 miliar dengan valuasi $852 miliar, tetapi COO Brad Lightcap dan CRO Denise Dresser akan pergi, dengan COO Wiz Dali Rajic mengambil alih penjualan.
- → OpenAI reportedly completed a $7 billion employee tender offer
- → OpenAI lets employees cash out another $7 billion in stock
- → Another OpenAI executive takes off
- → Brad Lightcap, OpenAI’s longtime COO, is leaving to ‘start something new’
- → OpenAI is losing its second executive this week
- → OpenAI hires new CRO as executive shake-up continues
Perkakas Agen & Aplikasi Lokal
Aplikasi model lokal. Unsloth merilis aplikasi desktop sumber terbuka untuk menjalankan dan melatih model lokal di berbagai GPU, dengan fitur eksekusi kode bersandbox, RAG, dan ekspor model.
Agen mengakses API web. Pratinjau pengembang Cloudflare memungkinkan situs web mengekspos alat MCP ke agen AI melalui sakelar dasbor, dan pelacakan agen baru menambahkan span untuk pemanggilan, panggilan model, dan eksekusi alat ke jejak Workers.
Itulah ulasan minggu ini - sampai jumpa Minggu depan.