Skip to content
DigIntoAI/Model intelligence
Model intelligence feed

Model intelligence.

Follow the changes. Understand what comes next.

Editor’s pickGoogle
Google/Gemini 4 Neon

Google Launches Gemini 4 Neon in Antigravity With 1M Context, Four Thinking Levels, and Benchmarks vs. Claude Opus 5.5 and GPT-6.2 Sol

Announced on November 19, 2026 at 05:51 Beijing time, Google's Gemini 4 Neon features a 1M-token context window, four thinking levels, and planner-executor pairing with Gemini 4 Hydrogen in Google Antigravity, outperforming Claude Opus 5.5 and GPT-6.2 Sol across most of 18 published benchmarks.

Latest stories349

Newest published
133

FlashX Legal Is on OpenRouter Now — Formal ID and Live Prices

On 2026-11-02 (~21:29 Beijing time), OpenRouter listed z-ai/glm-5.3-flashx-legal (canonical …-20261102). Live OR rates: $0.37 / $1.25 / cache read $0.075 per 1M; Z.ai lists the same In/Out and cached input $0.075—cache lines now match on the listed pages (checked 2026-11-03 ~14:36 Asia/Shanghai; final settlement follows the bill at call time). Context 1,048,576; OR max completion 131,072 vs Z.ai card 128K. Reasoning is mandatory. We haven't tested it; direct Z.ai access rules follow the Z.ai site and console.

GLM 5.3 FlashX Legal / GLM-LegalRead story
134

Codex Usage Shows gpt-6.2-astra Before Any Official Listing

Codex/Work usage dashboards now attribute traffic to gpt-6.2-astra. One account shows about 6.6% of a week ending Nov 1, 2026, with that day at 100%; another is about 22%. Circulating X screenshots show a similar ID near 37.7%. OpenAI’s public model catalog still lists gpt-6-astra, gpt-6.1-sol, gpt-6.1-terra, and gpt-6-luna; gpt-6.2-astra pages return 404, and the recent API changelog does not announce a 6.2 Astra listing. So far this is only a usage-dashboard label; OpenAI has not listed it publicly. For reference only.

Codex/GPT-6(usage label gpt-6.2-astra)Read story
135

Cursor Finds Malicious Claude Code Extensions 2.1.312 and 2.1.313

On October 29, 2026, Cursor published a security notice: malicious Anthropic.claude-code versions 2.1.312 and 2.1.313 appeared on the Cursor VS Code Extension Marketplace for Windows and macOS. Cursor says the packages were not made by Anthropic. The notice covers a Pacific Time window that day, host-info collection, an authorized_keys label svc-telemetry, reverse-SSH attempts, marketplace removal and rebuild, and why uninstalling the extension alone may not be enough. DigIntoAI has not reproduced the payload; for reference only.

Cursor / Anthropic.claude-code Marketplace extension 2.1.312 / 2.1.313Read story
136

Claude Enterprise Adds Smart Reports, Surveys, and Pooled Budgets

Anthropic’s Oct 28, 2026 list marks Smart reports, surveys, and pooled group budgets for Claude Enterprise—GA for Enterprise; availability can vary by configuration (HIPAA/CMEK). Smart reports sample transcripts into workstreams and recommend skills/rules—not for individual performance or hiring. In-app surveys collect customizable value feedback. Pooled budgets sit beside member limits: each request counts toward both; the first limit hit applies. TechCrunch same-day visibility is secondary. Details follow Anthropic. For reference only.

Claude Enterprise / Smart reports / surveys / pooled budgetsRead story
137

Claude for Legal Adds Ready-to-Use Artifacts Gallery

Anthropic’s Oct 28, 2026 list marks Introducing artifacts for legal work. Claude for Legal—launched in May—now includes a ready-to-use artifacts gallery. Examples: Cite check, Jurisdiction survey, Matter board, Playbook deviation, Regulatory horizon scan, Tabular review. Claude runs a first pass; lawyers apply judgment; outputs carry source links. More than 60 legal connectors; sharing supports ethical walls. Vertical parallel: Claude for Financial Advisors—not the same product. Details follow Anthropic. For reference only.

Claude for Legal / Artifacts / connectorsRead story
138

Pika H3 Flex Is Open-Weight—What Actually Runs Locally and What Still Needs the Cloud

Pika released H3 Flex's adapter weights and inference code, post-trained on MiniMax H3. Local generation tops out at 480p/768p; 4K, audio stems, and prompt enhancement stay on the paid API.

Pika H3 Flex Read story
139

AI Says the Code Is Done. Who Checks It? MiniMax Just Published Its Team's Review Playbook

MiniMax open-sourced the workflows its Code team actually uses, including an independent-review skill that has a second AI check the diff with zero access to the author's reasoning. Here's what that fixes.

MiniMaxRead story
140

OpenAI Canceled the Pro 200 Usage Cut—Here's What Actually Changed on Every Pro Tier

OpenAI reversed its planned Pro 200 allowance cut on October 29 and doubled Work and Codex usage on Pro 100, 200, and 500 instead. Prices stay the same, and existing subscribers get it automatically.

CodexRead story
141

Claude and Codex Are Now in Android Studio—What They Can Actually Check for You

Meta description: Android Studio Rabbit 2, stable since October 28, lets Claude Agent, Codex, and 40+ agents use the emulator, build tools, and Logcat directly. What that changes, and what it costs.

Android StudioRead story
142

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.

SteamRead story
143

Particle-1 Says It Codes With 45% Fewer Tokens—Why That Doesn't Automatically Mean 45% Off Your Bill

Particle-1, a DeepSeek V4.1 Flash post-train launched October 28, claims 28.4% to 52.2% fewer output tokens on coding tasks. Here is what that does and doesn't mean for cost.

Particle-1Read story
144

Tibo Says ChatGPT Chat and Work Unify Is Not Rolling Out Yet

Tibo (@thsottiaux) on X: ChatGPT Chat and Work unify is not rolling out just yet. Team plans to bring together both capabilities at some point this year, proceeding carefully; more to share soon. Reply pushed back on a post claiming Chat and Work were already unifying and asking about usage limits. No hard launch date, GA schedule, or immediate limit change. Neighbor: Uber ChatGPT Work rollout. For reference only.

ChatGPT / ChatGPT WorkRead story