US AI Brief — 2026-06-27

Posted on June 27, 2026 at 08:23 PM

US AI Brief — 2026-06-27

Top Stories


1. Gemini 3.5 Pro Delayed to July 2026

  • Build Fast with AI · 2026-06-26
  • Summary: Alphabet has confirmed that Gemini 3.5 Pro will not achieve public general availability in June 2026, missing Google CEO Sundar Pichai’s commitment made at Google I/O on May 19. The delay is attributed to “final adjustments” needed for the model. Prediction markets now price the odds of a June 30 launch at approximately 50%.
  • Why It Matters: The delay compounds Google’s challenges following the talent exodus and extends the window where competitors like OpenAI and Anthropic maintain advantage in frontier model capabilities. It also raises questions about Google’s ability to execute on its AI roadmap.
  • URL: AI News Today June 26 2026: 15 Biggest Stories

2. US Commerce Lifts Export Block on Anthropic’s Mythos 5

  • Indian Express / Markets Day · 2026-06-27
  • Summary: U.S. Commerce Secretary Howard Lutnick has lifted the export block on Anthropic’s Mythos 5 AI model, approving its release to approximately 100 trusted U.S. companies and agencies. The move follows a two-week standoff during which Anthropic disabled both Mythos 5 and Fable 5 after national security concerns were raised about the model’s autonomous offensive cybersecurity capabilities.
  • Why It Matters: This partial resolution signals a new paradigm of “controlled AI rollout” where frontier models are released only to vetted entities. It establishes a template for how the US government may manage future advanced AI deployments, balancing innovation with security.
  • URL: Anthropic resumes limited Mythos 5 rollout after US eases restrictions

3. GPT-4.5 Officially Retires from ChatGPT

  • OpenAI / Yahoo Tech · 2026-06-27
  • Summary: OpenAI has officially retired GPT-4.5, its last remaining GPT-4 generation model, from ChatGPT following a 30-day sunset period announced in late May. Existing conversations that used GPT-4.5 will continue with GPT-5.5. The reasoning model o3 is scheduled for retirement on August 26, 2026, following a 90-day transition.
  • Why It Matters: This marks the complete transition from GPT-4 to GPT-5 architecture in ChatGPT, reflecting OpenAI’s rapid model iteration cycle. Developers and enterprises relying on GPT-4.5 must now migrate to newer models, accelerating adoption of the latest AI capabilities.
  • URL: OpenAI just quietly retired the final GPT-4 model from ChatGPT

4. Google Loses 4 Senior AI Researchers to Anthropic in One Week

  • AIToolsRecap / Build Fast with AI · 2026-06-26
  • Summary: Google DeepMind experienced its worst talent exodus in history as four senior researchers departed in seven days: John Jumper (Nobel Prize 2024, AlphaFold), Jonas Adler (AI coding lead), Alexander Pritzel (pretraining specialist), and Arthur Conmy (Gemini 2.5, AI safety) all joined Anthropic. Additionally, Noam Shazeer (Transformer co-author, Gemini co-lead) joined OpenAI. Alphabet lost over $270 billion in market cap across two trading sessions.
  • Why It Matters: This talent drain represents a significant competitive shift in the AI race, with Anthropic consolidating top research talent. The departures highlight internal challenges at Google DeepMind and raise questions about Google’s ability to maintain its AI leadership position.
  • URL: AI News June 26 2026: Google Loses 4 AI Researchers in One Week