Anthropic releases Claude Opus 4.1 for enterprise with extended agent memory

Read the original article →

What happened

Anthropic released Claude Opus 4.1 on August 11 for enterprise customers. The update adds extended agent memory, improved long-horizon planning and what Anthropic described as a 30% reduction in hallucination rates on internal agent benchmarks.

The model is available through Claude Enterprise, Amazon Bedrock and Google Vertex AI. Anthropic said Opus 4.1 completed the White House voluntary frontier review window before release.

Why it matters

Long-horizon agents need memory that persists across sessions without blowing context limits. Opus 4.1 targets that gap for enterprise workflows like research, compliance review and multi-day coding tasks.

Passing the new frontier review process sets a precedent for how future Anthropic releases will be timed.

MintedBrain take

Test memory features on your longest real workflows, not demo threads. Extended memory increases token cost; pair Opus 4.1 with the spend controls Anthropic shipped in July.

References

This article was originally published at Amazon Bedrock. For the full piece, read the original article.

Discussion

  • Loading…

← Back to News