As the competition among AI models continues to intensify, Anthropic has released the latest version of Sonnet – this company's mid-range model. Anthropic indicates that the new version of Sonnet will run significantly faster than its predecessor, and also at a much lower cost.

The laboratory describes Sonnet 5.5 as an ideal assistant for daily tasks, including coding and creating office documents.

The predecessor to Sonnet, Sonnet, was released about three months ago. At that time, the selling point of this model was its efficient agentic deployment capability, which meant it could run agents at a cost lower than that of its competitors.

The main selling point of version 5.5 is speed. Anthropic claims that Sonnet version 5.5 is 30% faster than its predecessor, and token its energy consumption is also significantly slower.

In the model system of Anthropic, the capabilities of Sonnet are not as strong as those of Opus, but due to its greater flexibility, it may be more useful in certain situations. In particular, the benchmark tests of Anthropic show that Sonnet 5.5 performs better than Opus 5.5 in terms of agentic coding. This is likely because it is able to generate multiple agents without exceeding cost constraints.

Anthropic also indicates that Sonnet 5.5 possesses strong network security capabilities, and the company claims it is 'comparable' to Opus 5. Based on this, Anthropic states that 5.5 will be the first model to implement the same network security protection measures as Fable and Opus.

The company also plans to launch a new version of Haiku in the coming weeks – which is its smallest model – although no specific date has been given.

Over the past year, the main AI laboratories have been intensively releasing new models. Just last week, OpenAI released several new models, including enhanced versions of their mid-range and low-cost models Sol and Luna. Meta also announced a new model and stated that it will support an upcoming feature related to their smart glasses.