Why Are AI Companies Releasing New Models After Calling for AI to Slow Down?

24 September 2026 | By the Supra Vista Team

▶

Just 10 days after leading AI executives publicly backed calls to slow the pace of frontier AI development, Anthropic and OpenAI released new models.

Anthropic launched Claude Opus 5.5, while OpenAI released GPT-6 Sol and GPT-6 Luna on 22 September.

So are AI companies saying one thing and doing another?

“Slow down” does not necessarily mean “stop”

Anthropic CEO Dario Amodei has called for “pacing the frontier”—slowing how quickly the most powerful AI capabilities advance so safety measures and testing have more time to keep up. Other AI leaders, including OpenAI CEO Sam Altman, subsequently supported slowing the pace of frontier-model development.

Anthropic describes Opus 5.5 as its first model released since that call. The company says it underwent external safety evaluations before launch and delivers stronger performance than Opus 5 while costing about 40% less to run.

OpenAI says GPT-6 Sol and Luna bring capabilities from its larger GPT-6 family into faster, more affordable models. Its published API pricing confirms substantially lower costs than its top-tier Astra model.

The real tension: safety versus competition

This suggests an important distinction.

AI companies can try to slow advances at the frontier while continuing to improve efficiency, cost and existing capabilities.

But competition has not disappeared. Anthropic, OpenAI, Google, xAI and other labs are still racing for users, developers and businesses.

For families, it is a useful reminder that AI development is not simply about making technology more powerful. How quickly it advances—and how safely—is becoming part of the conversation too.

Explore how Supra Vista helps families understand and use AI responsibly →