Is Day 17 gone? Did Tibo fall asleep and forget?
Model intelligence feed
Model intelligence.
Follow the changes. Understand what comes next.
Editor’s pickOther providers
Valve's Answer to AI Slop and Clone Games on Steam: We Won't Police Developers' Tools
At Steam Indie Connect in Korea, Valve staff addressed AI slop and clone games head-on: no gatekeeping on AI tools, and the best defense against clones is community and steady updates.
Latest stories341
Newest published193194195196197198199200201202203204
Japan's Domestic LLMs Come in Three Routes
DigIntoAI maps Japan's domestic LLMs into three editorial routes—not an official taxonomy: closed API models, open-weight models, and Japan-focused post-training. Keep the Digital Agency's March 2026 源内 shortlist separate from later flagships, and do not treat Sakana Namazu as the same data-sovereignty tier as SoftBank Alloy. We have not independently benchmarked these models.
Japan's domestic LLMs (three editorial routes)Read story
How Qoder Takes Tasks from WeCom After You Set Up the Plugin
On October 22, 2026, Qoder's Chinese release notes say that after you finish the matching plugin setup, you can chat with Qoder in WeCom (企业微信), start tasks, and get replies—so work chat and task handoff stay in one thread. That is not the same as a chatbot that only talks. We have not tested the flow; plan and fees are not verified from the notes.
Qoder WeCom (企业微信) integrationRead story
What MiniMax Code Mini App Is and Where It Won't Run
MiniMax Code Mini App is a long-running plugin with a frontend, a Node backend that can keep data, and two-way calls with the Agent. It runs only in local macOS and Windows desktop sessions—not in cloud, SSH, or Remote Control. We have not tested it, and Mini App plan fees are not verified from the public pages.
MiniMax Code Mini AppRead story
Ant Group's Ming-3.0-tts Ranks Near the Top on Artificial Analysis Open TTS Boards
Artificial Analysis says Ant Group Ming-3.0-tts tops open TTS boards as the only sub-1B model in its top three. We have not tested.
Ming-3.0-tts (0.5B / tiny)Read story
Claude Code 2.1.309 Fixes Sandbox Gaps and Requires Trust for Ultrareview
Claude Code 2.1.309 patches sandbox, MCP and Edit-deny gaps; ultrareview now needs workspace trust. We have not tested.
Claude Code 2.1.309Read story
Anthropic Used Claude to Formally Verify 123 Open-Source Libraries
Anthropic says Claude proved no-crash specs for 239 OSS-Fuzz entry points in 123 libraries at RISC-V assembly. We have not tested the proofs.
Claude / formal verification (Frontier Red Team + Theorem)Read story
Anthropic-Managed Claude on Google Cloud Is Generally Available
Anthropic-managed Claude on Google Cloud is GA. Anthropic is the data processor; day-zero Platform features with IAM and billing. We have not tested.
Claude Models on Google Cloud (Anthropic managed)Read story
Claude Auto Mode Will Be Default on Web, Desktop and Mobile Next Week
Claude will default Pro, Max and Team on web, desktop and mobile to auto mode next week. Auto checks block risky actions; We have not tested.
Claude auto mode (web / desktop / mobile)Read story
OpenAI Decisions API Is Generally Available with GPT-6 Luna
OpenAI's Decisions API is GA on GPT-6 Luna for classification, routing and scoring. OpenAI claims up to 10× vs Luna on Responses; We have not tested.
Decisions API / GPT-6 LunaRead story
Cline Adds Free Access to Beam, Nemotron 3.5 and Three More Open Models
Cline said five open-weight models are now free in the app, including Beam and Nemotron 3.5. Terminal-Bench claims are Cline's; We have not tested.
Cline free models (Beam, Nemotron 3.5 Super VL, Laguna S 2.1, Nemotron 3.5 Lightning, Solar Mini 4)Read story
Google Cloud Introduces Agent Compute Engine for Isolated Agent Workloads
Google Cloud launched Agent Compute Engine, a hardware-isolated microVM for agents. Preview in coming months; Agent Substrate is open-source on GKE.
Agent Compute Engine / Agent SubstrateRead story
Anthropic Donates Agent Skills to the Agentic AI Foundation
Anthropic donated Agent Skills to the Agentic AI Foundation, the same home as MCP. Anthropic says more than 70 platforms integrated since Dec 2025.
Agent Skills / Agentic AI Foundation (AAIF)Read story
From the Webmaster
Personal experiences, thoughts and notes.