Claude Sonnet 4.5 (anthropic.com)

Article summary

Anthropic has released Claude Sonnet 4.5, a new coding model that offers improved performance and capabilities, including enhanced reasoning and math abilities. The model is available through the Claude API and is compatible with various tools and platforms. Claude Sonnet 4.5 is designed to be more aligned with human values and has undergone extensive safety training to reduce concerning behaviors. The model is being released under the AI Safety Level 3 (ASL-3) protections.

Main themes

• AI coding models
• Claude Sonnet 4.5
• AI safety and alignment
• Coding performance and capabilities
• Model comparisons and benchmarks

What commenters say

• Some users find Claude Sonnet 4.5 to be more efficient and effective for coding tasks, but others prefer Gemini 2.5 Pro for its ability to handle long contexts and complex codebases.
• The choice of AI model depends on the specific use case and task requirements, with different models exceling in different areas.
• Context management is a significant challenge when working with AI coding models, and some users prefer models that can handle longer contexts or have better context management features.
• Some users have had negative experiences with Claude's tendency to diverge from the planned task or produce overly complicated code, while others have found it to be a reliable and efficient tool.
• Gemini 2.5 Pro is considered to be particularly strong in handling long contexts and complex codebases, but may have issues with context blending and second-guessing the user.
• The use of tools like LiteLLM, OpenRouter, and AI SDK can simplify the process of working with multiple AI models and APIs.
• Some users prefer to use GPT-5-Codex for its ability to produce complex and detailed code, while others find it to be overly verbose and prone to errors.
• Effective context management strategies, such as compacting or clearing the context, are essential for getting the most out of AI coding models.