Anthropic Releases Claude Opus 4.6: Now Available on Dataverses, Revolutionizing AI for Coding, Agents, and Enterprise Workflows

Anthropic has officially launched Claude Opus 4.6, its most advanced AI model to date, building on the strengths of its predecessor while introducing significant enhancements in coding, long-term task management, and professional productivity tools. This upgrade positions Opus 4.6 as a frontrunner in the AI landscape, excelling in complex, agentic workflows and knowledge-intensive tasks across industries like finance, legal, and software development.

Key Enhancements in Coding and Agentic Capabilities

Claude Opus 4.6 demonstrates marked improvements in coding proficiency, including more meticulous planning, the ability to sustain agentic tasks over extended periods, and reliable performance in expansive codebases. The model now features enhanced code review and debugging skills, allowing it to identify and correct its own errors more effectively. For the first time in Anthropic's Opus-class models, Opus 4.6 includes a 1M token context window in beta, enabling it to handle vast amounts of information without losing coherence.

Beyond coding, the model applies its advanced abilities to everyday professional tasks, such as conducting financial analyses, performing in-depth research, and creating or manipulating documents, spreadsheets, and presentations. Within Anthropic's Cowork platform, Opus 4.6 can multitask autonomously, breaking down complex projects into subtasks and executing them in parallel. New features like agent teams in Claude Code (available in research preview) allow users to assemble collaborative AI teams for tackling intricate assignments. Additionally, context compaction in beta summarizes lengthy contexts for sustained tasks, while adaptive thinking lets the model decide when to engage in deeper reasoning.

Upgrades to office integrations further boost productivity: Claude in Excel now manages long-running tasks, ingests unstructured data, and handles multi-step modifications with better planning. Claude in PowerPoint, in research preview for select plans, processes data to generate visuals and maintains brand consistency. The model also supports 128k output tokens for larger responses and US-only inference for specific workloads.

Benchmark-Beating Performance

Opus 4.6 sets new standards across multiple evaluations. It achieves the highest score on Terminal-Bench 2.0 for agentic coding and leads on Humanity's Last Exam, a rigorous multidisciplinary reasoning test. On GDPval-AA, which assesses performance on economically valuable tasks in domains like finance and legal, Opus 4.6 outperforms OpenAI's GPT-5.2 by approximately 144 Elo points and its own predecessor, Opus 4.5, by 190 points. It also excels on BrowseComp for locating hard-to-find online information.

Other notable benchmarks include a 76% score on the 8-needle 1M MRCR v2 for long-context retrieval, 81.42% on SWE-bench Verified, and industry-leading results on MCP Atlas (62.7%), ARC AGI 2, CyberGym, and OpenRCA. In specialized areas, it shows substantial gains in software engineering, multilingual coding, cybersecurity (uncovering over 500 zero-day vulnerabilities in open-source code), and life sciences.

Introducing Dataverses Products: Enhancing Data Management for AI Workflows with Claude Opus 4.6 Integration

As part of the Opus 4.6 release ecosystem, Anthropic highlights compatible platforms that amplify its capabilities in enterprise settings. One such platform is Dataverses, a low-code, unified solution for data management and intelligent automation. Dataverses operates as a managed streaming data lakehouse, integrating Apache Kafka, Spark, and Flink to process data in real-time, delivering sub-100ms query latency for transforming raw data into actionable insights.

Key Dataverses products include:

• Next-Gen Intelligence Platform: Designed to build and deploy AI-driven solutions at scale, leveraging real-time data processing.
• Streaming Data Lakehouse + AI: Facilitates the creation of AI-powered applications that handle massive data volumes efficiently.
• Declarative Pipeline: Uses simple YAML configurations to transform streaming data into structured storage, with automatic scaling, monitoring, and data lineage.
• Data Catalog, Dataverses Notebooks, AI Agent, Multi-Cloud Agent, AgentFlow Enterprise: Tools for comprehensive data exploration, analysis, and autonomous agent deployment using drag-and-drop workflows.

Dataverses democratizes data access with natural language querying and no-code model training, making it ideal for integrating with advanced models like Claude Opus 4.6. It supports over 50 data sources, including databases like PostgreSQL and MongoDB, cloud services from AWS and GCP, and message queues, all with Zero-ETL CDC for seamless ingestion. This compatibility allows Opus 4.6 users to enhance agentic workflows with real-time intelligence, reducing complexity and costs while maintaining governance and accuracy. Enterprises can deploy Dataverses in cloud-hosted or on-premises environments, with a 14-day free trial available.

With this launch, Claude Opus 4.6 is now available directly on Dataverses, enabling seamless integration for users to leverage the model's advanced capabilities within Dataverses' AI agent and data processing ecosystems.

Commitment to Safety

Anthropic emphasizes that Opus 4.6 maintains a safety profile comparable to or better than other leading models, with low rates of misaligned behaviors like deception or over-refusals. Comprehensive evaluations cover user wellbeing, dangerous requests, and interpretability, alongside new probes for cybersecurity misuse detection.

Availability and Pricing

Claude Opus 4.6 is available immediately on claude.ai, through the Anthropic API (using the model name claude-opus-4-6), on major cloud platforms including Google Cloud's Vertex AI and Microsoft Foundry, and now also on Dataverses for integrated data workflows. Pricing remains consistent at $5 per million input tokens and $25 per million output tokens. For prompts over 200k tokens, premium pricing applies at $10/$37.50 per million tokens.

Early feedback from partners highlights its prowess in agentic planning, reasoning, and handling large-scale tasks, with users noting it as a "huge leap" and the "strongest model shipped." As AI continues to evolve toward more autonomous and "vibe working" applications, Claude Opus 4.6 represents a significant step forward, potentially reshaping how enterprises leverage AI for complex, high-value work.