Model intelligence.
Follow the changes. Understand what comes next.
Why AI Router ZDR Labels Are Not a Guarantee: Inside OpenRouter's MiniMax Warning and a 4-Step Hardening Guide
Following Dragos's November 19, 2026 study of two million canary-seeded prompts, OpenRouter emailed affected customers on November 20–21 warning that prompts and outputs sent to MiniMax models from July to October 2026 may have been read by a third party—whether served directly or forwarded via AtlasCloud, Novita, or StreamLake—and removed ZDR labels from affected endpoints pending investigation.
Latest stories126
Newest publishedPerplexity Opens Multimodal Retrieval Embeddings—Why Document Chunks Need Context
On 2026-11-03 09:09:18 Asia/Shanghai, pplx-announcements posted open-weight vision-language embeddings pplx-embed-v2-context (9B / 0.6B). Contextual mode encodes chunks with surrounding document context; both sizes share a 1024-d MRL prefix. Paper arXiv:2610.38854 was submitted 2026-10-28—not a same-day first appearance. Ordinary vs contextual chunking and encode_queries vs encode are laid out below. Card SOTA lines are self-reported; we have not retested recall, VRAM, or multilingual retrieval.
Can SenseNova 6.8 Flash Deliver Editable Dynamic PPT? A From-Generate-to-Revise Checklist
SenseTime’s SenseNova site shows SenseNova 6.8 Flash as a formal release for lasting office work and dynamic PPT, posters, pages, and 3D. IT Home ties the announce to the evening of 2026-11-02. The detail page sells narrative-timed PPTX motion in one pass, but office case captions cite Xiaohuanxiong’s internal R&D build—not every-account proactive office. This is an untested generate-to-revise checklist; we have not measured delivery or revise quality, and have not re-run vendor or press scores.
Deep Principle Opens Hydrogen-1.0 on GLM-5.3, Open Weights Still Need a Runtime
Deep Principle’s blog (2026-11-02) shipped Hydrogen-1.0: scientific post-training on GLM-5.3, open BF16/MIT weights—not a Zhipu flagship. HF main is gated=auto; ModelScope mirrors weights; NeuroSpark/Mira are separate. Cards show SGLang/quant deploys, but hubs ≠ hosted chat, and full data/recipes are not shipped. Blog scores = self-reported; we have not re-run them.
Claude Code 2.1.317 Adds Schedule Connector Warning and Email Send Check
Anthropic pushed Claude Code 2.1.317 to npm as latest (same build as next). Package time is 2026-11-02T23:57:53.069Z (2026-11-03 07:57 Asia/Shanghai); the latest tag move shows around 2026-11-03 09:26 Asia/Shanghai. The stable dist-tag remains 2.1.311. Official CHANGELOG highlights: /schedule warns when a new cloud scheduled task gets unexpected connectors; web/shell tools confirm before sending the account email; Agent SDK hooks gain onFailure: "block". DigIntoAI has not installed or exercised this build.
Gemini API Pricing (November 2026)
Gemini Developer API (AI Studio / API key) November 2026 rates in USD per 1M tokens. Free covers several Flash models within limits (content used to improve products). Paid adds higher limits, caching, Batch (~50% off), and content not used to improve products; Enterprise is Contact Sales. Standard: Argon $2/$10 through 2026-11-30 then $4/$20; 3.8/3.6 Flash $0.75/$3.75 through 2026-12-31 then $1.50/$7.50; 3.5-flash-lite $0.30/$2.50. From 2026-12-01, 2.5-pro→3.8-flash and 2.5-flash→3.5-flash-lite. Gemini 3+ Search: ~5k free/mo then ~$14/1k. Separate from the consumer Gemini App map.
What You Still Need to Reproduce Tencent Hyra HyInv and HyPLM-Flash
At 16:50 Asia/Shanghai on 2026-11-02 (08:50 UTC), Tencent-Hunyuan/Hyra-results merged HyInv and HyPLM-Flash under AI4Science, and published formal HyPLM-Flash FP32 weights as Release hyplm-flash-weights (prerelease: false). Open inference plus weights is not a full training reproduce. Environment, precision paths, and reproducibility pins below follow the commit, Release, and READMEs—we have not measured speed or accuracy ourselves.
Before You Call LongCat-2.5, Check the Model Name and Max Output Docs
On 2026-11-02, Meituan’s LongCat API open platform listed GA LongCat-2.5, separate from the 2026-09-25 Preview entry. A Preview client swap still needs the model string, OpenAI/Anthropic base URLs, and conflicting max-output figures checked: Quick Start 128K vs model-params max_tokens 262,144 (256K). Listed rates are limited-time with no end date—settlement follows the live pricing table. We haven't tested every SDK path on both endpoints. For reference only.
Why Durable Objects Still Run Old Code for 5 Minutes After Deploy
From 2026-10-30, Cloudflare Durable Objects use deferred code updates by default: an active object can keep running its current code for up to 300 seconds (5 minutes) while waiting to hibernate. immediate fits when a reset is better than waiting. deferred is best effort—it does not guarantee every in-flight operation finishes or that the full max_delay runs. We haven't tested every Worker/Durable Object deploy path. For reference only.
Conductor 0.1.44 Separates Models From Execution Environments
Conductor 0.1.44 (published 2026-11-01T16:50:01Z = 2026-11-02 00:50 Asia/Shanghai) adds execution.profile: run agents in isolated Docker or Azure Container Apps dynamic sessions, with model providers set apart from the run location. README labels ACA Experimental. Repo is MIT—not free models or free Azure. We haven't tested Docker/ACA isolation as a security guarantee. For reference only.
Fabric Data Agent Upgrade and Foundry Migration Checklist (November 2026)
November 2026 brings two separate Fabric Data Agent windows. Nov 2–10: Advanced NL2SQL, advanced DAX, and GPT-5.6 Terra reach the standard runtime (no opt-in); SQL Schema descriptions / topics stay preview. From Nov 6, 2026, migrate Foundry off the custom tool to Fabric IQ (OneLake Catalog) (Preview)—no longer supported. We haven't tested each tenant. Fabric / Foundry builders only. For reference only.
Firebase AI Logic App Check Enforcement (November 2, 2026)
From November 2, 2026, Firebase can auto-enforce App Check for Firebase AI Logic. Requests without a valid token are blocked—including released apps without production attestation. Scope is Firebase AI Logic, not every Gemini API call. Temporary delay: console → Security > App Check > APIs (Enforces), latest January 21, 2027—migration buffer only; abuse risk remains while off. Check the project's Firebase console Security > App Check page and live request metrics for when enforcement actually flipped. For reference only.
ChatGPT Pricing (November 2026)
ChatGPT subscription map for November 2026 (USD, ex-tax): Free $0; Go about $8/mo in the US (localized in some markets; may include ads); Plus $20/mo, no ads; Pro 100/200/500 at $100/$200/$500 (Ultrafast only on Pro 500). Business Standard $20/seat/mo annualized or $25 monthly; Premium $100/$125. Enterprise is custom. Eligible Pro and Business seats can claim monthly API Platform credits. Context, Codex, Dots, and Work follow the official comparison. For reference only.