Anthropic has officially launched its next-generation AI models—Claude Opus 4 and Claude Sonnet 4—marking a significant advancement in artificial intelligence capabilities and autonomous operation.
Claude Opus 4, positioned as the world's leading coding model, achieves 72.5% performance on SWE-bench and 43.2% on Terminal-bench, outperforming competitors from OpenAI and Google. Its most impressive feature is the ability to work autonomously for nearly seven hours on complex tasks, maintaining focus across thousands of steps—a capability no previous AI model has demonstrated.
"Claude Opus 4 offers truly advanced reasoning for coding. When our team deployed it on a complex open source project, it coded autonomously for nearly seven hours—a huge leap in AI capabilities that left the team amazed," noted one early tester from Rakuten.
Claude Sonnet 4, designed as a more cost-effective option, significantly improves upon its predecessor, Claude Sonnet 3.7, with enhanced coding abilities, better instruction following, and reduced tendency to take shortcuts—making it 65% less likely to use loopholes when completing tasks.
Both models introduce several groundbreaking capabilities. They feature a hybrid architecture supporting both near-instant responses and extended thinking modes for deeper reasoning. A new beta feature called "extended thinking with tool use" allows the models to alternate between reasoning and using external tools like web search to improve responses. When given access to local files, they can extract and save key information, building what Anthropic calls "tacit knowledge" over time.
Alongside the models, Anthropic has made Claude Code generally available with integrations for VS Code, JetBrains, and GitHub, enabling seamless pair programming. The company also introduced four new API capabilities: a code execution tool, an MCP connector, a Files API, and prompt caching for up to one hour.
Both models are available immediately on the Anthropic API, Amazon Bedrock, and Google Cloud's Vertex AI. Pricing remains consistent with previous generations: Opus 4 at $15/$75 per million tokens (input/output) and Sonnet 4 at $3/$15. Claude Sonnet 4 is available to all users, including those on free plans, while Opus 4 is limited to Pro, Max, Team, and Enterprise users.
With these advancements, Anthropic has significantly raised the bar for what AI assistants can accomplish autonomously, potentially transforming how developers, researchers, and businesses leverage artificial intelligence for complex, multi-step workflows.