Willison's April 2026 newsletter: Opus 4.7, GPT-5.5 price hikes, Claude Mythos security

Simon Willison’s sponsors-only April newsletter covers the month’s frontier model churn: Anthropic’s Opus 4.7 and OpenAI’s GPT-5.5 both shipped with price increases, and ChatGPT Images 2.0 landed alongside a wave of smaller model releases. The issue also digs into Claude Mythos and ongoing LLM security research, plus Willison’s monthly roundup of tools he’s actually using.

The newsletter sits behind a $10/month sponsorship that runs a month ahead of the public archive, with the March edition available as a free preview. Adjacent recent posts on the blog include a major backwards-compatible refactor of LLM 0.32a0, a retrospective on the now-defunct OpenAI/Microsoft AGI clause, and analysis of DeepSeek V4 closing in on frontier capability at a fraction of the cost — a recurring theme of price pressure cutting in both directions across the model landscape.