Sonnet 5.5 (anthropic.com)

Article summary

Anthropic has released Claude Sonnet 5.5, a faster and lower-cost model that complements Claude Opus 5.5. Sonnet 5.5 improves over its predecessor in performance, collaboration, cost, and speed, making it suitable for everyday tasks and coding. The model has shown gains in multiple areas, including agentic coding and knowledge work. It is priced the same as Sonnet 5, but typically needs fewer tokens to complete tasks, resulting in lower costs.

Main themes

• AI model performance
• Cost and pricing
• Safety and security
• Cyber Verification Program
• Restrictions and gate-keeping
• Company priorities and values

What commenters say

• The new model's performance is comparable to Opus 5.5, but at a lower cost, making it a more attractive option for some users.
• The Cyber Verification Program is ineffective and does not provide the expected benefits, with some users experiencing flagging and degradation to earlier models.
• The safeguards in place are overly restrictive and hinder the ability to write secure code, leading some users to switch to open models.
• The company's approach to safety and security is seen as inconsistent and overly cautious, with some users feeling that it is not necessary or effective.
• Some users believe that the models are being unnecessarily restricted, and that fully unrestricted models could be a lucrative opportunity for the company.
• The documentation for the Cyber Verification Program is outdated and does not accurately reflect the current state of the program.
• The company's priorities and values are being questioned, with some users feeling that they are more focused on avoiding liability than on providing useful tools.
• The restrictions on the models are seen as a form of gate-keeping, with the company deciding what users can and cannot do with the models.