Reply to thread

Message

[QUOTE="Munyaradzi Mafaro, post: 38932, member: 636"]
Anthropic released Claude 4 artificial intelligence models that compete against OpenAI and Google. The company created two versions, Opus 4 and Sonnet 4, with better reasoning and coding abilities. Claude Opus 4 scored 72.5 percent on SWE-bench coding tests and beat all other AI models. Sonnet 4 works faster and also achieves strong results on the same difficult programming challenges. Both models can handle complex tasks that run for many hours without breaking down.

The new AI systems gained better memory and thinking skills for long projects. Cognition and Rakuten tested the models on demanding jobs that lasted seven hours each. Claude 4 makes fewer mistakes and avoids shortcuts compared to earlier versions. The models can work with tools and remember information from previous conversations. Because of these improvements, GitHub plans to use Sonnet 4 for its next coding assistant.

Anthropic also launched Claude Code for software developers working with popular programming tools. The service connects with JetBrains and VS Code editors to help programmers write better applications. Companies can access both models through Amazon Bedrock and Google Vertex AI platforms. Opus 4 costs 75 dollars per million tokens, and Sonnet 4 costs 15 dollars for the same amount.
[/QUOTE]

Name