Introducing Claude Opus 4.6

We’re upgrading our smartest model.

The new Claude Opus 4.6 improves on its predecessor’s coding skills. It plans more carefullysustains agentic tasks for longercan operate more reliably in larger codebasesand has better code review and debugging skills to catch its own mistakes. Andin a first for our Opus-class modelsOpus 4.6 features a 1M token context window in beta¹.

Opus 4.6 can also apply its improved abilities to a range of everyday work tasks: running financial analysesdoing researchand using and creating documentsspreadsheetsand presentations. Within Coworkwhere Claude can multitask autonomouslyOpus 4.6 can put all these skills to work on your behalf.

The model’s performance is state-of-the-art on several evaluations. For exampleit achieves the highest score on the agentic coding evaluation Terminal-Bench 2.0 and leads all other frontier models on Humanity’s Last Exama complex multidisciplinary reasoning test. On GDPval-AA—an evaluation of performance on economically valuable knowledge work tasks in financelegaland other domains²—Opus 4.6 outperforms the industry’s next-best model (OpenAI’s GPT-5.2) by around 144 Elo points,³ and its own predecessor (Claude Opus 4.5) by 190 points. Opus 4.6 also performs better than any other model on BrowseCompwhich measures a model’s ability to locate hard-to-find information online.

As we show in our extensive system cardOpus 4.6 also shows an overall safety profile as good asor better thanany other frontier model in the industrywith low rates of misaligned behavior across safety evaluations.

Knowledge workAgentic searchCodingReasoning

In Claude Codeyou can now assemble agent teams to work on tasks together. On the APIClaude can use compaction to summarize its own context and perform longer-running tasks without bumping up against limits. We’re also introducing adaptive thinkingwhere the model can pick up on contextual clues about how much to use its extended thinkingand new effort controls to give developers more control over intelligencespeedand cost.

We’ve made substantial upgrades to Claude in Exceland we’re releasing Claude in PowerPoint in a research preview. This makes Claude much more capable for everyday work.

Claude Opus 4.6 is available today on claude.aiour APIand all major cloud platforms. If you’re a developeruse claude-opus-4-6 via the Claude API. Pricing remains the same at $5/$25 per million tokens; for full detailssee our pricing page.

We cover the modelour new product updatesour evaluationsand our extensive safety testing in depth below.

First impressions

We build Claude with Claude. Our engineers write code with Claude Code every dayand every new model first gets tested on our own work. With Opus 4.6we’ve found that the model brings more focus to the most challenging parts of a task without being told tomoves quickly through the more straightforward partshandles ambiguous problems with better judgmentand stays productive over longer sessions.

Opus 4.6 often thinks more deeply and more carefully revisits its reasoning before settling on an answer. This produces better results on harder problemsbut can add cost and latency on simpler ones. If you’re finding that the model is overthinking on a given taskwe recommend dialing effort down from its default setting (high) to medium. You can control this easily with the /effort parameter.

Here are some of the things our Early Access partners told us about Claude Opus 4.6including its propensity to work autonomously without hand-holdingits success where previous models failedand its effect on how teams work:

Claude Opus 4.6 is the strongest model Anthropic has shipped. It takes complicated requests and actually follows throughbreaking them into concrete stepsexecutingand producing polished work even when the task is ambitious. For Notion usersit feels less like a tool and more like a capable collaborator.

Early testing shows Claude Opus 4.6 delivering on the complexmulti-step coding work developers face every day—especially agentic workflows that demand planning and tool calling. This starts unlocking long-horizon tasks at the frontier.

Claude Opus 4.6 is a huge leap for agentic planning. It breaks complex tasks into independent subtasksruns tools and subagents in paralleland identifies blockers with real precision.

Claude Opus 4.6 is the best model we've tested yet. Its reasoning and planning capabilities have been exceptional at powering our AI Teammates. It's also a fantastic coding model – its ability to navigate a large codebase and identify the right changes to make is state of the art.

Claude Opus 4.6 reasons through complex problems at a level we haven't seen before. It considers edge cases that other models miss and consistently lands on more elegantwell-considered solutions. We're particularly impressed with Opus 4.6 in Devin Reviewwhere it's increased our bug catching rates.

Claude Opus 4.6 feels noticeably better than Opus 4.5 in Windsurfespecially on tasks that require careful exploration like debugging and understanding unfamiliar codebases. We’ve noticed Opus 4.6 thinks longerwhich pays off when deeper reasoning is needed.

Claude Opus 4.6 represents a meaningful leap in long-context performance. In our testingwe saw it handle much larger bodies of information with a level of consistency that strengthens how we design and deploy complex research workflows. Progress in this area gives us more powerful building blocks to deliver truly expert-grade systems professionals can trust.

Across 40 cybersecurity investigationsClaude Opus 4.6 produced the best results 38 of 40 times in a blind ranking against Claude 4.5 models. Each model ran end to end on the same agentic harness with up to 9 subagents and 100+ tool calls.

Claude Opus 4.6 is the new frontier on long-running tasks from our internal benchmarks and testing. It's also been highly effective at reviewing code.

Claude Opus 4.6 achieved the highest BigLaw Bench score of any Claude model at 90.2%. With 40% perfect scores and 84% above 0.8it’s remarkably capable for legal reasoning.

Claude Opus 4.6 autonomously closed 13 issues and assigned 12 issues to the right team members in a single daymanaging a ~50-person organization across 6 repositories. It handled both product and organizational decisions while synthesizing context across multiple domainsand it knew when to escalate to a human.

Claude Opus 4.6 is an uplift in design quality. It works beautifully with our design systems and it’s more autonomouswhich is core to Lovable’s values. People should be creating things that matternot micromanaging AI.

Claude Opus 4.6 excels in high-reasoning tasks like multi-source analysis across legalfinancialand technical content. Box’s eval showed a 10% lift in performancereaching 68% vs. a 58% baselineand near-perfect scores in technical domains.

Claude Opus 4.6 generates complexinteractive apps and prototypes in Figma Make with an impressive creative range. The model translates detailed designs and multi-layered tasks into code on the first trymaking it a powerful starting point for teams to explore and build ideas.

Claude Opus 4.6 is the best Anthropic model we’ve tested. It understands intent with minimal prompting and went above and beyondexploring and creating details I didn’t even know I wanted until I saw them. It felt like I was working with the modelnot waiting on it.

Both hands-on testing and evals show Claude Opus 4.6 is a meaningful improvement for design systems and large codebasesuse cases that drive enormous enterprise value. It also one-shotted a fully functional physics enginehandling a large multi-scope task in a single pass.

Claude Opus 4.6 is the biggest leap I’ve seen in months. I’m more comfortable giving it a sequence of tasks across the stack and letting it run. It’s smart enough to use subagents for the individual pieces.

Claude Opus 4.6 handled a multi-million-line codebase migration like a senior engineer. It planned up frontadapted its strategy as it learnedand finished in half the time.

We only ship models in v0 when developers will genuinely feel the difference. Claude Opus 4.6 passed that bar with ease. Its frontier-level reasoningespecially with edge caseshelps v0 to deliver on our number-one aim: to let anyone elevate their ideas from prototype to production.

The performance jump with Claude Opus 4.6 feels almost unbelievable. Real-world tasks that were challenging for Opus [4.5] suddenly became easy. This feels like a watershed moment for spreadsheet agents on Shortcut.

01 / 20

Evaluating Claude Opus 4.6

Across agentic codingcomputer usetool usesearchand financeOpus 4.6 is an industry-leading modeloften by a wide margin. The table below shows how Claude Opus 4.6 compares to our previous models and to other industry models on a variety of benchmarks.

Opus 4.6 is much better at retrieving relevant information from large sets of documents. This extends to long-context taskswhere it holds and tracks information over hundreds of thousands of tokens with less driftand picks up buried details that even Opus 4.5 would miss.

A common complaint about AI models is “context rot,” where performance degrades as conversations exceed a certain number of tokens. Opus 4.6 performs markedly better than its predecessors: on the 8-needle 1M variant of MRCR v2—a needle-in-a-haystack benchmark that tests a model’s ability to retrieve information “hidden” in vast amounts of text—Opus 4.6 scores 76%whereas Sonnet 4.5 scores just 18.5%. This is a qualitative shift in how much context a model can actually use while maintaining peak performance.

All in allOpus 4.6 is better at finding information across long contextsbetter at reasoning after absorbing that informationand has substantially better expert-level reasoning abilities in general.

Long-context retrievalLong-context reasoning

Finallythe charts below show how Claude Opus 4.6 performs on a variety of benchmarks that assess its software engineering skillsmultilingual coding abilitylong-term coherencecybersecurity capabilitiesand its life sciences knowledge.

Root cause analysisMultilingual codingLong-term coherenceCybersecurityLife sciences

A step forward on safety

These intelligence gains do not come at the cost of safety. On our automated behavioral auditOpus 4.6 showed a low rate of misaligned behaviors such as deceptionsycophancyencouragement of user delusionsand cooperation with misuse. Overallit is just as well-aligned as its predecessorClaude Opus 4.5which was our most-aligned frontier model to date. Opus 4.6 also shows the lowest rate of over-refusals—where the model fails to answer benign queries—of any recent Claude model.

For Claude Opus 4.6we ran the most comprehensive set of safety evaluations of any modelapplying many different tests for the first time and upgrading several that we’ve used before. We included new evaluations for user wellbeingmore complex tests of the model’s ability to refuse potentially dangerous requestsand updated evaluations of the model’s ability to surreptitiously perform harmful actions. We also experimented with new methods from interpretabilitythe science of the inner workings of AI modelsto begin to understand why the model behaves in certain ways—andultimatelyto catch problems that standard testing might miss.

A detailed description of all capability and safety evaluations is available in the Claude Opus 4.6 system card.

We’ve also applied new safeguards in areas where Opus 4.6 shows particular strengths that might be put to dangerous as well as beneficial uses. In particularsince the model shows enhanced cybersecurity abilitieswe’ve developed six new cybersecurity probes—methods of detecting harmful responses—to help us track different forms of potential misuse.

We’re also accelerating the cyberdefensive uses of the modelusing it to help find and patch vulnerabilities in open-source software (as we describe in our new cybersecurity blog post). We think it’s critical that cyberdefenders use AI models like Claude to help level the playing field. Cybersecurity moves fastand we’ll be adjusting and updating our safeguards as we learn more about potential threats; in the near futurewe may institute real-time intervention to block abuse.

Product and API updates

We’ve made substantial updates across ClaudeClaude Codeand the Claude Platform to let Opus 4.6 perform at its best.

Claude Platform

On the APIwe’re giving developers better control over model effort and more flexibility for long-running agents. To do sowe’re introducing the following features:

• Adaptive thinking. Previouslydevelopers only had a binary choice between enabling or disabling extended thinking. Nowwith adaptive thinkingClaude can decide when deeper reasoning would be helpful. At the default effort level (high)the model uses extended thinking when usefulbut developers can adjust the effort level to make it more or less selective.
• Effort. There are now four effort levels to choose from: lowmediumhigh (default)and max. We encourage developers to experiment with different options to find what works best.
• Context compaction (beta). Long-running conversations and agentic tasks often hit the context window. Context compaction automatically summarizes and replaces older context when the conversation approaches a configurable thresholdletting Claude perform longer tasks without hitting limits.
• 1M token context (beta). Opus 4.6 is our first Opus-class model with 1M token context. Premium pricing applies for prompts exceeding 200k tokens ($10/$37.50 per million input/output tokens)available only on the Claude Platform.
• 128k output tokens. Opus 4.6 supports outputs of up to 128k tokenswhich lets Claude complete larger-output tasks without breaking them into multiple requests.
• US-only inference. For workloads that need to run in the United StatesUS-only inference is available at 1.1× token pricing.

Product updates

Across Claude and Claude Codewe’ve added features that allow knowledge workers and developers to tackle harder tasks with more of the tools they use every day.

We’ve introduced agent teams in Claude Code as a research preview. You can now spin up multiple agents that work in parallel as a team and coordinate autonomously—best for tasks that split into independentread-heavy work like codebase reviews. You can take over any subagent directly using Shift+Up/Down or tmux.

Claude now also works better with the office tools you already use. Claude in Excel handles long-running and harder tasks with improved performanceand can plan before actingingest unstructured data and infer the right structure without guidanceand handle multi-step changes in one pass. Pair that with Claude in PowerPointand you can first process and structure your data in Excelthen bring it to life visually in PowerPoint. Claude reads your layoutsfontsand slide masters to stay on brandwhether you’re building from a template or generating a full deck from a description. Claude in PowerPoint is now available in research preview for MaxTeamand Enterprise plans.