AI News · Models ·
Anthropic unveils Claude Opus 5.5 as labs introduce cheaper models
Anthropic unveiled Claude Opus 5.5 on September 22. CNBC reported the release as part of a push toward less costly models amid competition from open-weight alternatives.
Key points
- Anthropic unveiled Claude Opus 5.5 on September 22.
- Anthropic said Opus 5.5 will cost around 40% less to run than Opus 5.
- OpenAI announced GPT-6 Sol and Luna, with API prices 50% below GPT-5.6 promotional pricing.
- CNBC linked the launches to customer spending concerns and competition from cheaper, open-weight models.
- Independent quality comparisons and workflow-specific savings were not reported.
What happened: Anthropic unveiled Claude Opus 5.5 on September 22, joining OpenAI in announcing less costly artificial intelligence models as customers look to contain spending. CNBC reported that the launches reflect pressure on both AI developers from cheaper, open-weight competitors, including Alibaba, Moonshot AI and DeepSeek. For businesses choosing AI models, the announcements put operating costs alongside capabilities as a central consideration when deciding which model to use.
The details: Opus 5.5 is the latest release in Anthropic's Claude 5.5 model family. The company described it as a more token-efficient version that will cost around 40% less to run than Opus 5. Dianne Penn, Anthropic's head of product management, research and labs, told CNBC that the company is working to make thinking and answering more efficient, using fewer tokens depending on the effort setting. That distinction matters: Anthropic's claim concerns the cost of running the model, rather than an explicitly reported percentage reduction in its API prices. Exact pricing and workload-specific comparisons were not reported.
Background: OpenAI's parallel launch adds GPT-6 Sol and GPT-6 Luna to its GPT-6 family. CNBC reported that their API prices are 50% lower than GPT-5.6 promotional pricing. Sol sits below Astra and is intended for more complex work, including coding. Luna targets high-volume tasks such as extracting information and summarizing documents, OpenAI said. The two companies' savings figures therefore use different comparisons: Anthropic points to running costs against Opus 5, while OpenAI compares API prices with an earlier promotional rate.
Who it affects: The releases address customers seeking more cost-effective models and trying to rein in AI spending, CNBC reported. They also arrive during a debate over advanced AI safety. These are the first releases from either lab since Anthropic CEO Dario Amodei called for an industrywide slowdown in developing advanced AI. OpenAI CEO Sam Altman and Tesla and SpaceX CEO Elon Musk also joined that call, according to CNBC.
What to watch: For teams already using Opus, the announcement is a reason to revisit model selection, not assume that the newest version is automatically the best fit. Compare actual workflow costs and output quality, with attention to the effort settings behind Anthropic's efficiency claim. Independent quality comparisons and measured savings across specific business workflows were not reported. Those unanswered questions limit what buyers can conclude from the headline cost reduction alone.
Our take
Opus users have another reason to revisit their model selection. Compare actual workflow costs and output quality rather than assuming a newer model will be the best fit.