Anthropic's Opus 5.5 is here
What's the story
Anthropic has launched its latest AI model, Opus 5.5, which sets a new benchmark in coding and knowledge work performance. The new model is cheaper than its predecessor, with output tokens costing $20 per million tokens for Opus 5.5 as opposed to the previous model's $25 price tag. Other metrics have also seen similar price reductions, making the new version more economical to run while requiring less compute power overall.
Enhanced communication
Major improvements in communication
The release of Opus 5.5 comes just two months after the launch of its predecessor, Opus 5.
The new model has made major improvements in how it communicates, using less jargon and placing key information at the start of messages.
Anthropic also claims that Opus 5.5 outperforms the larger Fable model in many benchmarks and informal tasks where Fable failed to deliver results.
Advanced features
On par with Mythos in biology and cybersecurity capabilities
Anthropic has said that Opus 5.5 is on par with Mythos in biology and cybersecurity capabilities.
However, its release comes with the same safeguards as the company's Fable model, which restricts its use for discovering exploits in compiled programs or developing recognizable biological weapons.
This is part of Anthropic's commitment to AI safety and responsible development practices.
Safety measures
Commitment to AI safety
The safety training for Opus 5.5 was similar to its predecessors, with alignment testing and pre-release evaluation by external organizations like METR and Frontier Design. Anthropic has said that more advanced training and evaluation systems are already being prepared for future models.
Market launch
What about its availability?
Opus 5.5 has been described by Anthropic as a major upgrade from its predecessor, with improvements in agentic coding, computer use, and knowledge work.
The model is now available across platforms such as Amazon Web Services, Google Cloud, and Microsoft Azure.
This broad availability is part of Anthropic's strategy to make advanced AI technology accessible to a wider audience, while continuing to prioritize safety and responsible development practices.