Last Week in AI #345 - 5 new models, 9 misalignment incidents, some Dots

Key points

• Anthropic launches Claude Opus 5.5 with stricter safeguards for cybersecurity1
• OpenAI launches GPT-6.1 Sol, says it nearly matches GPT-6 Astra and costs less1
• First, Anthropic announced Claude Opus 5.5, which runs 40 percent cheaper than Opus 5 while matching Fable 5.1 on most work, and inheriting Fable-style safeguards: cybersecurity requests get re-routed to the weaker Opus 4.8, and flagged biology requests go to Opus 5.1
• OpenAI, meanwhile, extended its GPT-6 generation with updated Sol and Luna models, released 90 minutes after Anthropic’s Opus 5.5 update.1

Sources (1)

• [1]Last Week in AI #345 - 5 new models, 9 misalignment incidents, some Dots
  Last Week in AI · Sep 30, 11:23 AM

  Anthropic and OpenAI race to release smarter and cheaper models

  - Anthropic launches Claude Opus 5.5 with stricter safeguards for cybersecurity

Extractive summary: sentences quoted from the sources.