Anthropic has introduced Claude Opus 4.1, an enhanced version of its flagship Claude Opus 4 model. Released on August 5, 2025, Opus 4.1 delivers noticeable performance improvements in software engineering, agentic reasoning, and research-level tasks—yet remains a seamless upgrade via the same pricing and API footprint as its predecessor.
Key features and updates include:
- Improved coding accuracy: Opus 4.1 scores 74.5% on SWE‑bench Verified, surpassing Opus 4’s 72.5% . It does so through finer-grained multi-file refactoring, more precise bug fixes, and context-aware style adaptation.
- Sharper research and agentic reasoning: The model excels at long-horizon, tool-assisted tasks—synthesizing insights across external data sources and executing complex workflows with improved detail orientation .
- Hybrid reasoning & flexible control: As a “hybrid reasoning” model, Opus 4.1 supports both rapid responses and step-by-step thinking. Developers can now fine-tune “thinking budgets” via API for cost-efficiency and performance balance .
- Safety assurance: Despite being an incremental update, Anthropic conducted safeguarded evaluations to confirm the model’s risk profile remains consistent with Opus 4. The model preserves reliability while demonstrating marginally better refusal rates on harmful requests .
- Easy upgrade path: Existing Opus 4 users can continue with minimal changes. The model is accessible via
claude-opus-4-1-20250805 in the API, available to paid Claude users (Pro, Max, Team, Enterprise), Claude Code subscribers, and through Amazon Bedrock and Google Cloud’s Vertex AI. Pricing remains unchanged from Opus 4 .
Real-world feedback:
- At Rakuten, teams reported that Opus 4.1 “pinpointed the exact spot requiring correction—without making unnecessary adjustments or introducing new bugs” .
- Windsurf observed a performance gain equivalent to the jump from Claude Sonnet 3.7 to Sonnet 4—a full standard deviation advantage over Opus 4 .
In context:
Claude Opus 4.1 follows quickly on the May 22, 2025 launch of Claude 4 models (Opus 4 and Sonnet 4), which introduced extended reasoning, tool use, and sustained coding capabilities . Opus 4 itself was already a standout for tasks like seven-hour independent coding runs and enterprise-level workflows . With Opus 4.1, Anthropic delivers a performance tuning that’s targeted, efficient, and straightforward to adopt.