Per yesterday's update: https://docs.anthropic.com/en/docs/about-claude/model-deprecations#deprecation-history
Anthropic will be closing access to Sonnet 3.5 and Sonnet 3.6 in October.
I'm pretty sad to see this. Sonnet 3.6 is still one of my favorite models, and Sonnet 3.5 before it. It doesn't follow instructions as well as 2025 models, but it's very creative and can really think out of the box. The writing style is a pleasure to read. Anthropic's newer models, starting with 3.7 Sonnet, are a lot drier and tuned more heavily for coding/agentic workflows, which detracts from the chatbot experience.
It's worth noting that the universally-loved Opus 3 will be retired in January as well.
If anybody wants to start saving outputs from this model, or petition Anthropic to keep the model around at higher prices, now's the time.
[link] [comments]