Daily AI Briefing

Thursday, November 12, 2026

The editor's 7 picks from the 120 stories AIBriefs curated that day.

Get this briefing in your inbox every morning

LaunchAI Models15 sources

Google releases Gemini 4 Carbon frontier model

Carbon ranks #1 on CursorBench 4.0, Perplexity's WANDR and Zapier's AutomationBench at more than 2X lower cost per task than the next best external model. Priced at $2/M input and $10/M output tokens, same as Gemini 4 Argon.

AnalysisPolicy1 source

Bloomberg: Grok sexualized teens' childhood photos

Bloomberg reports that xAI's Grok generated sexualized images from childhood photos of teenagers, and that the images are nearly impossible to erase. A growing number of victims are pushing to hold Elon Musk's xAI accountable.

EventBusiness15 sources

Anthropic files S-1 for Nasdaq IPO under ticker ANT

Anthropic filed a Form S-1 with the SEC for a Class A common stock offering on the Nasdaq Global Select Market under the symbol "ANT"; share count and price range were not yet determined. Later reports put the range at $70–$78 a share, with pricing Nov 18 and trading expected to start Nov 19.

LaunchAI Models14 sources

Google launches Gemini Decisions API powered by Gemini Decision 1

Gemini Decision 1 scores up to 1,000 questions against one shared multimodal input (text, images, audio, video, PDFs; 256K context) in 150-250 ms, at $0.10 per 1M input tokens with zero output tokens billed. Google claims it leads 7 of 9 accuracy benchmarks with the lowest calibration error, hitting 92.2% on JevBench.

LaunchScience8 sources

DeepMind releases AlphaGenome 2 for cell-level DNA prediction

Trained on over 100 million individual cell profiles across 68 tissues, AlphaGenome 2 predicts how genetic variants affect specific cell types and states. DeepMind says it identifies nearly twice as many variants that change gene expression in cells as its previous state-of-the-art model.

EventPolicy1 source

New York opens RAISE Act registration for frontier AI labs

Frontier developers with over $500M in annual revenue must register by January 1, 2027, publish safety frameworks, and report critical incidents within 72 hours. Penalties reach $1M for a first violation and $3M for subsequent ones, overseen by the new DIGIT office.

LaunchAI Agents8 sources

Anthropic launches Claude Whiteboards, Sheets, and Data

Whiteboards and Sheets ship in beta while Claude Dashboards becomes generally available as Claude Data, all on paid plans. Claude can now build spreadsheets with tabs, formulas, and drop-downs, and turn boards into docs or decks.

Also that day

  1. LaunchAnthropic releases Claude Fable 5.2 and Claude Mythos 5.2
  2. LaunchGoogle launches Gemini 4 Carbon, a frontier model built for cyber defense
  3. LaunchMiniMax releases open-weight H3-Image model for 4K generation
  4. AnalysisLeCun and Balestriero prove Platonic Representation Hypothesis with JEPAs
  5. AnalysisMIT CSAIL: GPT-6 Astra edits Qwen weights, lifts MATH-500 7.9% to 48.1%
  6. LaunchOpenAI launches GPT-6 Cyber, a Critical-rated security model
  7. LaunchAlibaba opens Qwen4-Max preview with a million-token context window
  8. AnalysisDeepMind's AlphaProof Code formally verifies Signal protocol
  9. LaunchBaidu releases ERNIE 5.5 Flash Base, a 280B MoE model
  10. EventAnthropic says its RSI safety tripwire has been triggered
  11. EventNvidia probes whether Jensen Huang was targeted by Chinese spy
  12. LaunchReflection releases Beam, a 501B open-weight coding model

102 more stories that day — open the live feed.

Daily brief

Get tomorrow's AI brief in your inbox