AI
Anthropic releases Claude Opus 5, positioning it as Fable 5's more affordable alternative
Weeks after government scrutiny and regulatory negotiations, Anthropic unveiled its latest model with enhanced cybersecurity protections and matched pricing against its predecessors.
The Verge
Anthropic announced the release of Claude Opus 5 on Thursday, its newest flagship model, arriving amid heightened regulatory oversight of AI systems and competitive pressure from rivals like OpenAI.
Performance and positioning#
The new model delivers capabilities approaching Claude Fable 5 across many domains while demonstrating particular strength in complex coding tasks. Anthropic positioned Opus 5 as the company's go-to offering for enterprise customers tackling knowledge work and biological research, while reserving Fable 5 for more ambitious applications and long-horizon AI agent projects.
According to the company, Opus 5 represents "the most aligned Opus model and is the least susceptible to being tricked into misuse," reflecting enhanced security measures integrated into its design.
Government coordination and safeguards#
Following weeks of negotiations between Anthropic and the Trump administration over AI safety concerns—which previously led to the temporary offline status of Fable 5 and Mythos 5—the new model incorporates substantially stronger cybersecurity protections than its predecessor, Opus 4.8.
When asked if the company submitted Opus 5 to government review before launch, Anthropic spokesperson Danielle Ghiglieri confirmed in an interview with The Verge: "We continue to work with our government partners to conduct their own independent testing of our models. This includes Opus 5."
Pricing and options#
Anthropic matched the per-token pricing of Opus 4.8 at $5 input and $25 output costs per million tokens, positioning Opus 5 as roughly half the price of Fable 5. The model undercuts OpenAI's GPT-5.6 on cost as well.
The company is introducing Fast mode as a research preview feature, allowing customers to prioritize speed at double the standard price. Users can also opt into automatic fallback mechanisms that route requests to cheaper models when Opus 5's safeguards decline certain tasks.
These pricing adjustments appear to respond to recent disputes over usage caps and concerns that AI developers are passing infrastructure costs to customers in opaque ways.
