Anthropic is testing Claude for Chrome, an AI that helps with tasks in the browser like managing calendars and emails. The pilot involves 1,000 trusted users to address safety risks, especially prompt injection attacks where malicious instructions trick the AI. New defenses cut attack success rates but more work is needed. Users control permissions and confirm risky actions. Feedback will improve safety and features before wider release.
Marketing Auditor simplifies your audit process, letting you generate comprehensive, white-label reports in just a few clicks. Save over 10 hours per report while analyzing 200+ data points and delivering 50+ pages of actionable insights. Customize reports with professional themes or your own branding, and export them in editable formats like PowerPoint or Google Slides to showcase your expertise.
Claude Opus 4.1 improves on Opus 4 with better agentic tasks, coding, and reasoning, scoring 74.5% on SWE-bench Verified. It excels in multi-file code refactoring and precise debugging, favored by users like Rakuten. Available to paid users, API, Amazon Bedrock, and Google Cloud, pricing is unchanged. Users are encouraged to upgrade and provide feedback as larger updates are planned.
Claude 4 introduces Claude Opus 4 and Sonnet 4, improving coding, reasoning, and AI agents. Opus 4 excels in complex tasks and coding, while Sonnet 4 enhances coding and instruction following. Both support tool use, parallel execution, and better memory with local file access. Claude Code is now available with IDE integrations and SDK. New API features enable powerful AI agents, reducing shortcut behaviors and improving task focus and safety.
Anthropic has launched a new subscription plan for its AI chatbot Claude called Max, priced at $100 and $200 per month. This plan offers higher usage limits than the $20 Claude Pro subscription and priority access to new features. The Max plan is designed for users needing extensive access for demanding projects, allowing up to 20 times more usage. It is ideal for those requiring extended conversations and quick responses. The plan is available now.
Claude can now search the internet for up-to-date responses, enhancing its accuracy with the latest information. It provides direct citations for fact-checking and delivers relevant sources in a conversational format. This feature helps sales teams, financial analysts, researchers, and shoppers by offering insights on trends, market data, grant proposals, and product comparisons. Web search is available for paid users in the U.S., with plans for broader access.
Anthropic is releasing Claude 3.7 Sonnet, a hybrid AI model that provides both real-time and thoughtful responses, allowing users to control its reasoning time. It is available through the Anthropic API, Amazon Bedrock, and Google Cloud's Vertex AI, with pricing starting at $3 per million input tokens. Claude 3.7 Sonnet excels in coding, visual data extraction, customer service, and content generation, and it demonstrates superior performance in various benchmarks.
Claude.ai now offers custom styles for users to tailor responses to their needs. Users can personalize Claude’s responses to match their communication preferences, tone, and structure. Options include Formal, Concise, and Explanatory styles. Users can also generate custom styles by uploading sample content and specifying instructions. Early adopters like GitLab have used styles to standardize communication and decision-making.
Claude.ai introduces an analysis tool for writing and running JavaScript code for data processing and real-time insights. Available in preview, it acts as a code sandbox for complex math and data tasks, enhancing accuracy and reproducibility. It supports teams by analyzing data, improving conversions, performance, and decision-making. Users can enable this feature by logging into Claude.ai and managing feature previews.
Anthropic announced upgraded Claude 3.5 Sonnet and new Claude 3.5 Haiku models. Claude 3.5 Sonnet shows significant coding and tool use improvements, while Claude 3.5 Haiku offers state-of-the-art performance at similar cost and speed. A new capability, computer use, allows Claude to interact with computers like humans, currently in public beta. Early adopters include Asana, Canva, and Replit. These models are available on Anthropic API, Amazon Bedrock, and Google Cloud.
