Anthropic and OpenAI each ship two models in one week, plus always-on Dots agents
Days after Amodei and Altman both said the industry needs to pace the frontier, Anthropic released Claude Opus 5.5 and Sonnet 5.5, OpenAI shipped GPT-6 Sol and Luna plus Dots agents that keep working after ChatGPT is closed, and GPT-6.1 Astra was pulled over safety concerns.
It took them only one week. So much for "pacing the frontier." Anthropic released two new models. OpenAI shipped two new models, and agents that work without you. At least they shelved GPT-6.1 Astra because it wasn't safe. Ten days before all of this, Amodei wrote that the industry has to slow down. Altman replied: "I agree with Dario that we need to pace the frontier." Then it all happened: Claude Opus 5.5 - flagship-level work, 40% cheaper to run than Opus 5 Claude Sonnet 5.5 - beats Opus 5.5 on agentic coding at half the price GPT-6 Sol and Luna - the GPT-6 family, faster and cheaper Dots - always on agents with their own computer, browser and 4,000+ apps, working after you close ChatGPT GPT-6.1 Astra - pulled before its October launch after safety concerns in internal testing To be fair, the top models stayed where they were. The ones right below them got cheaper, faster and more autonomous in the same week. For anyone building agents, this is the useful part. An agent that runs all day needs a model cheap enough to call thousands of times, and clear rules for what it can do alone. Both moved forward in seven days. Is this what pacing the frontier looks like to you?