Anthropic launches Claude Opus 4.1

Anthropic today has officially launched Claude Opus 4.1, its next-gen model with agentic tasks, real-world coding, and powerful reasoning. The model was spotted in the internal benchmarking file before making a superfast debut.

The Opus 4.1 has achieved 74.5% on the software engineering benchmark, compared to 72.5% achieved by Opus 4 and 62.3% by Sonnet 3.7. This new performance empowers the in-depth research and data analysis skills of the new model, especially for exploration and agentic tasks.

This new model is available for paid Claude users and in Claude Code. You can also access it via Anthropic’s API, Amazon Bedrock, and Google Cloud’s Vertex AI. All of these platforms offer the same price as Opus 4.

In addition to launching this new model, Anthropic has officially announced that the it will release larger improvemnets to its other models in the coming weeks.

Claude Opus 4.1 Benchmark (Source – Anthropic)