Anthropic Ships Opus 5.5 Model at 40% Lower Cost After Safety Debate

Anthropic released Opus 5.5 on Tuesday, marking the company's first new model in its Claude 5.5 family and its first launch since chief executive Dario Amodei publicly called for slowing the pace of frontier AI development.

The new model delivers performance comparable to Anthropic's Fable 5.1 while cutting costs by 40 percent compared to the previous Opus 5, according to the company. Opus 5.5 is priced at $4 per million input tokens and $20 per million output tokens, down from $5 and $25 respectively for Opus 5. Tokens represent small words or word fragments that AI models process.

Why it matters

The release demonstrates how AI companies are navigating competing pressures: advancing capabilities to remain competitive while addressing mounting safety concerns from within their own ranks. Pricing has emerged as a critical battleground as providers race to make powerful models economically viable for enterprise deployment.

Safety measures and internal tensions

Anthropic emphasized that Claude 5.5 models underwent extensive alignment testing, external evaluations, and safeguards for high-risk domains including cybersecurity and biology. The company stated it is developing new safety training scenarios and improving rewards for AI alignment work.

The launch follows public turmoil within the AI safety community. Earlier this month, Anthropic researcher Jacob Coxon announced his resignation on X, stating that neither Anthropic nor his previous employer OpenAI took AI safety seriously enough. Coxon wrote that developers "earnestly believe that it could kill us all by the end of the decade."

Evan Hubinger, Anthropic's alignment science lead, responded by confirming Coxon's assessment and stating he believes there is a greater than 10 percent chance the technology could "kill all humans."

Industry split on existential risk

Amodei subsequently published an essay advocating for slower AI development, drawing support from OpenAI CEO Sam Altman and SpaceXAI's Elon Musk. However, other industry leaders have rejected such warnings. Nvidia CEO Jensen Huang dismissed concerns about AI destroying humanity as unscientific, while President Trump has characterized AI fears as conspiracy theories.

The divergence highlights a fundamental split in how technology leaders assess the risks of increasingly capable AI systems, even as they continue shipping more powerful models to market.

Details of the Opus 5.5 release were first reported by Yahoo Finance.