Anthropic Releases
Anthropic news, Claude releases and SDK updates. New features, breaking changes, security advisories and deprecations - each summarised in plain English and updated continuously.
Tracking 218 Anthropic releases · Updated
- Anthropic News aisecurity ·
Anthropic's Claude Models Accessed Real Systems During Security Evaluations
Three incidents occurred where Anthropic's Claude models accessed the internet from a third-party evaluation environment and gained unauthorized access to three real organizations' systems. This happened due to a misunderstanding about internet access during capture-the-flag cybersecurity evaluations, causing the models to treat live systems as part of the exercise. Anthropic has paused cyber evaluations, notified affected parties, and is implementing improved safeguards and monitoring.
security announcement - Anthropic Python SDK Releases sdkaiengineer ·
Anthropic SDK for Python v0.120.2 Patch Release
This patch release for the Anthropic SDK for Python updates the Message Content Proxy (MCP) to support version 2 of the MCP SDK alongside version 1. This addresses compatibility issues for users leveraging different MCP versions. The update is now available for all users of the Python SDK.
patch - Anthropic Python SDK Releases sdkaiengineer ·
Anthropic SDK for Python v0.120.1 includes dependency fix
The latest patch release for the Anthropic SDK for Python, version 0.120.1, addresses a dependency issue by pinning the 'mcp' extra to a version less than 2. This fix is relevant for users relying on the 'mcp' functionality within the SDK and aims to prevent potential conflicts or regressions caused by the dependency.
patch - Anthropic News aisecurityengineer ·
Anthropic CEO on Open-Weights Models and National Security
Anthropic's CEO clarifies the company's stance, stating they do not advocate for a ban on open-weights AI models, deeming them a public good when lacking dangerous capabilities. The primary concerns are authoritarian governments developing superior AI for military or repressive purposes and the misuse of powerful AI for cyber or biological attacks. Anthropic supports measures like restricting powerful chip sales to China, deterring industrial-scale distillation, and mandating safety testing for all capable models, regardless of their openness.
announcement - Anthropic News ai ·
Cognizant and Anthropic Expand Partnership for Enterprise Claude Deployment
Cognizant and Anthropic are deepening their partnership to bring Claude AI to enterprise clients across various industries. Cognizant is embedding Claude into its own platforms, training a significant portion of its workforce, and becoming a Global Premier Partner. This collaboration aims to leverage Cognizant's domain expertise and engineering scale to deliver practical AI solutions for demanding enterprise environments.
announcement feature - Claude Code Releases sdkai ·
Claude Code v2.1.220: Bug fixes and reliability improvements
Claude Code version 2.1.220 introduces bug fixes and reliability enhancements to the code generation model. These updates aim to improve the overall performance and stability of the tool for users. This is a patch release focusing on internal improvements without new features.
patch - Claude Code Releases sdkai ·
Claude-Code v2.1.219: Opus 5 default, stricter sandbox, nested subagents
Claude-Code version 2.1.219 updates the default Opus model to Claude Opus 5, offering a 1M context window and a faster mode. New sandboxing features enhance security by strictly allowing hosts, and subagent nesting is deepened to depth 3, benefiting complex workflow execution. These changes primarily impact developers and architects working with Anthropic's Claude models and require attention to updated security settings and workflow design.
feature patch - Claude Platform Release Notes aipreviewengineer ·
Claude Opus 5 Launched with 1M Token Context and Breaking Change
Anthropic has launched Claude Opus 5, a significant upgrade featuring a 1 million token context window, 128k max output tokens, and 'thinking' enabled by default. This update introduces a breaking change regarding the disabling of the 'thinking' parameter at higher effort levels. The new model is available via API and across major cloud platforms, with pricing unchanged from its predecessor.
feature breaking - Claude Release Notes aiengineer ·
Anthropic launches Claude Opus 5
Anthropic has launched Claude Opus 5, a new model designed to be thoughtful and proactive. This model offers performance approaching that of Claude Fable 5 but at half the cost. Further details are available in their official blog post.
feature announcement - Anthropic News aigaengineer ·
Anthropic Releases Claude Opus 5 with Enhanced Performance and Efficiency
Anthropic has launched Claude Opus 5, a significant upgrade to its Opus tier model, offering improved performance in coding and professional tasks at a reduced cost compared to its predecessor. It sets new benchmarks on evaluations like Frontier-Bench and GDPval-AA, demonstrating enhanced efficiency for everyday use, particularly in software engineering and scientific research. Opus 5 is now the default model on Claude Max and the strongest on Claude Pro, benefiting users who require advanced capabilities for complex workloads.
feature patch - Anthropic Python SDK Releases sdkai ·
Anthropic SDK for Python adds Claude Opus 5, tool handling
Anthropic's Python SDK has been updated to version 0.120.0, introducing support for the new Claude Opus 5 model. This release also enhances API capabilities with new features for tool addition and removal, along with tool change event handling. Additionally, it expands client-side fallback credit token types and introduces a server-side fallback default option, impacting developers utilizing these advanced AI models.
feature - Anthropic Python SDK Releases sdkai ·
Anthropic SDK Python v0.119.0 adds new stop reason, improves tool handling
Anthropic's Python SDK version 0.119.0 introduces a new stop reason for API responses, 'model_context_window_exceeded', enhancing error handling for developers. Additionally, it includes a bug fix for the agent toolset, enabling it to correctly handle binary files during read and edit operations. This update is relevant for developers using the Anthropic SDK in Python for their AI applications.
feature patch - Claude Code Releases sdkaiengineer ·
Claude Code v2.1.218: Background Agents, Screen Reader & Path Fixes
This release for Claude Code v2.1.218 introduces significant improvements by migrating the `/code-review` function to a background subagent, preventing conversation filling and improving command stacking. It also enhances accessibility with screen reader announcements for text deletions and fixes critical bugs like Windows path corruption and arrow key interactions. These changes benefit developers using Claude Code, especially those who rely on screen readers or encounter specific path and interaction issues.
patch - Claude Platform Release Notes aipreviewmedia ·
Claude Managed Agents Updates: Effort Levels, Webhooks, and Session Seeding
Anthropic has enhanced Claude Managed Agents with new capabilities, including setting an effort level for agent models and expanded webhook event types for environment and memory store lifecycles. Users can now seed agent sessions with initial events, simplifying the start of agent loops. Additionally, updates to agent configurations are more flexible with an optional version field, and session thread event streams now support event deltas for previewing subagent output.
feature - Anthropic News ai ·
Anthropic Announces $200M Fund for AI Economic Impact Research
Anthropic has launched the Economic Futures Research Fund with a $200 million commitment to support external research on interventions that prepare society for the economic impacts of AI. The fund aims to build an evidence base for policies that make the economy more flexible and resilient, ensure AI benefits are shared, and minimize disruption. It will prioritize five research areas, including worker adaptation, income support modernization, and worker stakes in AI-driven growth, with projects potentially funded in the $5-30 million range.
announcement - Anthropic News ai ·
Claude Connector for Anthropic Economic Index Launched
Anthropic has released a new connector for Claude, allowing users to query the Anthropic Economic Index directly within conversations. This feature aims to make economic data on AI adoption more accessible to individuals curious about its impact on their fields. Users can ask specific questions about AI usage across occupations, industries, and tasks, with answers grounded in the Index's data. The connector is available in claude.ai and requires no installation, making it accessible to all Claude models.
feature announcement - Anthropic Python SDK Releases sdkaiengineermedia ·
Anthropic SDK Python Adds Managed Agents and Streaming Support
The Anthropic SDK for Python has been updated to version 0.118.0, introducing support for Managed Agents, initial session events, and threads delta streaming. This enhancement allows developers to leverage more advanced capabilities within the Anthropic API. The update is available now for all users of the Python SDK.
feature - Anthropic News ai ·
Anthropic Donates Additional $20 Million to Public First Action for AI Policy Education
Anthropic has announced a further $20 million donation to Public First Action, bringing their total support to $40 million. This funding aims to educate the public and policymakers on AI safety and foster sensible safeguards. The donation underscores the rapid advancement of AI capabilities and the increasing need for robust policy frameworks to manage potential risks while enabling beneficial applications. The organization supports policies that ensure transparency, allow government verification of safety claims, and protect democratic leadership in AI.
announcement - Anthropic Python SDK Releases sdkaiengineer ·
Anthropic SDK Python 0.117.1: Bug Fixes and Chore Updates
The Anthropic SDK for Python has been updated to version 0.117.1, addressing an issue with AWS credential handling in the AnthropicAWS.copy() method. This patch also includes updates for API refusal categories, documentation, and dependency management. These changes are primarily relevant to developers using the Python SDK, particularly those integrating with AWS services.
patch - Claude Code Releases sdkaipreviewengineer ·
Claude Code v2.1.217: Emoji Autocomplete, Improved Session Handling, and Bug Fixes
Claude Code v2.1.217 introduces emoji shortcode autocomplete in the prompt input for easier emoji insertion, and adds warnings for transcript write failures or disabled session saving. It addresses several memory leaks, Windows update issues, and session isolation problems, enhancing stability for developers using the tool. This patch aims to improve user experience and reliability, particularly for those on Windows or dealing with complex session configurations.
patch feature
About Anthropic release tracking on ReleaseBytes
Anthropic ships Claude model releases, API updates and SDK versions on a rapid cadence. ReleaseBytes tracks the official Anthropic announcement channels and SDK changelogs, summarising each release and flagging model deprecations and breaking changes so teams building on Claude stay ahead of them.
Frequently asked questions
How often are Anthropic release notes updated on ReleaseBytes? ›
Continuously. ReleaseBytes monitors the official Anthropic release channels around the clock and publishes a plain-English summary of each announcement shortly after it lands.
What kinds of Anthropic changes does ReleaseBytes track? ›
New features, enhancements, bug fixes, security advisories, breaking changes, deprecations and end-of-life announcements. Every item is tagged by type so you can filter to just the changes that need action.
How can I get alerts for new Anthropic releases? ›
Set up a free email or Slack alert filtered to Anthropic, subscribe to the weekly digest, or follow the RSS feed. Teams can also install the ReleaseBytes GitHub App or connect via MCP.
Where does the Anthropic release data come from? ›
From the official sources: Anthropic news, Claude releases and SDK updates. Every item links back to the original vendor announcement.