OpenAI and Anthropic Release New AI Models Amid Speed Control Debate

Anthropic and OpenAI released new artificial intelligence (AI) models within two hours on Tuesday, just 10 days after their CEOs publicly called for slowing AI development speed.
This time, both companies chose to lower prices instead of surpassing their strongest models. Elon Musk, who supported the AI speed control policy, also congratulated Anthropic.
Anthropic launched Claude Opus 5.5 at 16:31 UTC (4:31 PM GMT). The company stated that Opus 5.5 performs similarly to Claude Haiku 5.1 in most tasks while being 40% cheaper to run than Opus 5. The price reduction is based on Anthropic's standard benchmark tests, with list prices dropping 20%: $4 per million input tokens and $20 per million output tokens. Tokens are text segments that models read and write, which developers pay for by the token.
Opus 5.5 is the first model released since Dario Amodei, CEO of Anthropic, published an essay on September 12 urging a slowdown in AI development speed improvements. OpenAI CEO Sam Altman agreed with this stance, and Musk commented that 'Dario is right.' This was reported by BeInCrypto.
However, Amodei's essay left room for future product releases.
Anthropic revealed that external testing organizations METR and Frontier Design reviewed Opus 5.5 before its launch.
OpenAI responded with cheaper GPT-6 Sol and Luna models at 18:12 UTC (6:12 PM GMT). Both models are faster and more cost-effective than the base GPT-6 Astral, which remains OpenAI's flagship model.
OpenAI halved developer pricing for both models. Sol costs $2 per million input tokens, while Luna is priced at $0.10 per million output tokens—representing a 50% reduction compared to the promotional GPT-5.6 price. Paid ChatGPT users can immediately access both models, while free users can test Luna on desktop apps.
OpenAI and Anthropic have shifted the competitive landscape: a pricing war has begun. Critics had questioned the AI speed control pledge from the start. Investor Michael Burry called it an opportunistic exaggeration for IPO purposes, and Morgan Stanley, Anthropic's IPO lead manager, stated such moves would not affect AI-related costs.
Neither company has announced when they will release versions surpassing their current top models.
Korean Source
This article is an English localization of a Korean-language crypto news report. Original headline: AI 성장 둔화 끝? 오픈AI·앤트로픽 신규 모델 출시