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 publishedQwen Code 0.27.0 Lets Web Shell Users Ask for Details Before Choosing
Qwen Code 0.27.0 adds Ask for details to pending questions in the Web Shell main view when the daemon advertises session_side_task. A side task explains the options without answering or changing the main session. A separate update keeps eligible local child questions independent; unsent drafts do not survive refresh reliably.
How to Use Cursor Voice Chat Without Losing Track of Your Code
Some development problems are easier to explain aloud, especially when they involve a sequence of clicks or an idea you are still working out.
Claude Code 2.1.319 Adds Background MCP Tasks and Fixes Security Issues
Anthropic released Claude Code 2.1.319 with terminal MCP task support, a setting that denies calls instead of waiting for permission prompts, and security fixes. npm latest and next point to 2.1.319; stable remains 2.1.311. DigIntoAI hasn't installed or exercised this build.
Which Offline Phone Tasks Fit Alibaba's MNN-Pocket-0.3B
Alibaba's MNN-Pocket-0.3B is a 300M-parameter Chinese/English instruction model for on-device assistants, extraction and lightweight agent pipelines. Its INT4 W4A16 QAT MNN package is about 173MB of deployment files, not runtime RAM, with 32K supported context and Apache 2.0 licensing. The publisher reports about 350 decode tok/s on iPhone 18 Pro; actual app reliability and resource use still need device-level evaluation.
Codex CLI 0.167.0 Adds Agent Command Center and Cross-Backend Switching
OpenAI shipped Codex CLI 0.167.0 (GitHub rust-v0.167.0; 2026-11-06T01:40:00Z / 2026-11-06 09:40 Asia/Shanghai; npm latest 0.167.0). Highlights: codex agents --no-daemon agent command center with unread markers; cross-backend switching that keeps drafts; local image previews; optional gcp_setup_wizard (setup wizard, not first Vertex); /experimental system proxy; expand patch during approval. Brief fixes/docs/chores. We haven't tested these flows here. Latest GitHub Release wording applies. For reference only.
GitHub Copilot Pricing (November 2026)
GitHub Copilot subscription map for November 2026 (USD, ex-tax): Free $0 with limited chat/agent and 2,000 completions/month; Student free for verified students with unlimited completions; Pro $10/mo with 1,500 AI credits; Pro+ $39 with 7,000; Max $100 with 20,000. Business $19/seat/mo with 1,900 pooled credits/user; Enterprise $39/seat/mo with 3,900; Copilot access has no included allowance. Completions on paid plans do not consume credits. Credits reset at 00:00 UTC on the 1st of each calendar month. For reference only.
Grok API Text Watermark Starts Nov 7—How to Check
xAI’s Output Watermarking docs (Last updated October 31, 2026) say that starting November 7, 2026—about two days from 2026-11-05 (Asia/Shanghai)—listed Grok 4.7/4.6/4.5/4.3 and 4.20-series models add a default invisible watermark on API text to comply with the EU AI Act. Coverage includes Responses and Chat Completions: ordinary text, structured outputs, and tool call arguments. x-watermark: true means that reply was marked; any other value or a missing header means it was not. Only Team admins can opt out per model in Console. Docs give no clock time or time zone for November 7.
Step 5 Preview Cache Hit Price Doubles Nov 6—Why Your Total Bill Might Not
Tencent Cloud's TokenHub & CloudBase notice raises Step 5 Preview (step-5-preview) cache-hit input from 2026-11-06 00:00:00 Asia/Shanghai: Guangzhou 0.35→0.70, Singapore 0.36→0.72 RMB/M tokens. Miss input and output stay unchanged (numbers not listed). As of 2026-11-04, still a preview—not yet live. A doubled cache-hit line ≠ a doubled total bill; impact tracks hit share. TokenHub/CloudBase only, not a StepFun all-channel hike. DigIntoAI has not checked bills or hit rates—figures follow the Tencent Cloud notice.
AWS Distills GLM-5.3 Into a Qwen3.5-4B Shopping Agent—Near the Teacher, With Tradeoffs
AWS Machine Learning Blog (02 NOV 2026; Zappia et al.) shows SeqKD of a shopping agent: Bedrock GLM-5.3 teacher → ShoppingBench rejection sampling → LoRA SFT on Qwen3.5-4B (SageMaker AI). Held-out 324 problems, single LoRA run: student ASR 0.541 vs greedy teacher 0.540 vs base 0.297; tool-only 0.415; IFEval ≈6–8 pts down. No per-task cost/latency published. AWS-reported; we haven't retested.
Claude Code 2.1.318 Adds Login Expiry Warnings and Stops Auto-Installing IDE Extensions
Anthropic pushed Claude Code 2.1.318 to npm as latest (same as next). Package time 2026-11-03T21:55:17.247Z (2026-11-04 05:55 Asia/Shanghai); modified and GitHub Release v2.1.318 land around 2026-11-04 08:01 Asia/Shanghai. stable stays 2.1.311. CHANGELOG: loginExpiresAt / credential_expires_at plus a 10-day stderr warning; IDE terminals no longer auto-install or update the IDE extension (/ide or the editor extensions view). DigIntoAI has not installed or exercised this build.
Claude Pricing (November 2026)
Claude.ai subscription map for November 2026 (USD, ex-tax): Free $0; Pro $20/mo or $17/mo with $200 annual prepay; Max $100/mo; Ultra $200/mo. Comparison table: Free Limited/Limited on 5-hour and weekly windows; Pro 1x/1x; Max 5x/5x; Ultra no 5-hour limit for verified accounts / 10x weekly. Max includes $100/mo API credits; Ultra $200/mo. Team Standard $20/$25 per seat; Premium $100/$125; seats mixable for 2–150 people; pooled Team API credits cap $500/mo. Enterprise US$20/seat/mo annual with 20-seat minimum plus usage at API rates. Separate from the Claude API model wall. For reference only.
How Do You Prove an API Didn't Swap Models? Anthropic Opens FlashProve
On 2026-11-02, Anthropic’s Frontier Red Team published Progress on provable inference for large language models and open-sourced FlashProve: a zkSNARK research prototype for Commit→Prove→Verify. A provider can commit to weights without publishing them, then attach a checkable proof that a reply came from those committed weights. The proof covers model identity and the inference path—not factual correctness. Official Maverick figures are self-reported; we have not retested. This is not a Claude API feature already live for all users.