OpenAI and Anthropic launch cheaper models after slowdown calls

Anthropic and OpenAI launched new AI models less than two hours apart on Tuesday. Both companies cut prices instead of pushing past their most powerful models. Anthropic released Claude Opus 5.5 at 16:31 UTC. Anthropic said Claude Opus 5.5 matches Claude Fable 5.1 on most tasks. Anthropic said Claude Opus 5.5 costs 40% less to run than Opus 5. Anthropic's own tests measured the 40% reduction on typical workloads. Anthropic cut list prices by 20%. Claude Opus 5.5 costs $4 per million input tokens. Claude Opus 5.5 costs $20 per million output tokens. Anthropic CEO Dario Amodei published an essay on September 12 urging AI labs to slow capability gains. Amodei wrote that pacing does not mean halting model training or technical progress. He wrote that companies should take adequate time to align and safeguard their models. OpenAI CEO Sam Altman agreed with the call. Elon Musk wrote that Amodei was right. Musk also congratulated Anthropic. Anthropic said outside testers METR and Frontier Design reviewed Claude Opus 5.5 before release. OpenAI introduced GPT-6 Sol and GPT-6 Luna at 18:12 UTC. OpenAI said both models are faster and lower-cost models built on GPT-6 Astra. GPT-6 Astra remains OpenAI's flagship. OpenAI halved prices for developers. GPT-6 Sol costs $2 per million input tokens. GPT-6 Luna costs $0.10 per million input tokens. The 50% cut is measured against GPT-5.6 promotional pricing. The cut is not measured against standard rates. Paid ChatGPT users get both models today. Free users can try GPT-6 Luna in the desktop app. Investor Michael Burry called the slowdown pledge self-serving IPO hype. Morgan Stanley said the push changes nothing for spending. Neither lab has said when it will ship a model above its current flagship.