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 125 Anthropic releases · Updated
- Anthropic Python SDK Releases sdkaiengineer ·
Anthropic SDK for Python v0.109.1 fixes API refusal category
Anthropic's Python SDK has been updated to version 0.109.1, introducing a bug fix for its API. This update adds the 'frontier_llm' refusal category, providing clearer information on specific API rejection reasons. This change is relevant for developers using the SDK to interact with Anthropic's language models.
patch - Claude Code Releases sdkaigaengineer ·
Anthropic Claude Code v2.1.170: Fable 5, Session Fixes
Anthropic has released Claude Code version 2.1.170, introducing the new Mythos-class model Claude Fable 5, now safe for general use and exceeding previous model capabilities. This update also resolves an issue where sessions and transcripts were not saved or retrievable via `--resume` when launched from specific terminal environments. This patch is available now for all users.
feature patch - Claude Release Notes ai ·
Anthropic Launches Claude Fable 5 Model
Anthropic has released Claude Fable 5, a Mythos-class model that is now safe for general use. This launch provides a new, more broadly applicable AI model for users. Further details are available in a related blog post.
feature announcement - Claude Platform Release Notes aipreviewengineer ·
Anthropic Launches Claude Fable 5 and Mythos 5 with 1M Token Context
Anthropic has released Claude Fable 5, their most capable widely available model, and Claude Mythos 5 for Project Glasswing participants. Both models feature a 1 million token context window and enhanced adaptive thinking capabilities. These updates impact users interacting with the Claude API, particularly concerning tokenizer changes, safety classifier responses, and data retention policies for Fable 5.
feature announcement - Anthropic News ai ·
Anthropic Launches Claude Fable 5 and Mythos 5 AI Models
Anthropic has released Claude Fable 5, a highly capable model for general use with safety safeguards, and Claude Mythos 5, a version with lifted safeguards for cybersecurity professionals. Fable 5 excels in software engineering, knowledge work, and vision tasks, while Mythos 5 demonstrates advanced capabilities in drug design and scientific hypothesis generation. Both models are priced significantly lower than their predecessors and are available with different access models.
feature announcement - Claude Code Releases sdkaiengineerfinance ·
Claude Code v2.1.169: Troubleshooting, Directory Navigation, and UI Fixes
This update to Claude Code introduces a `--safe-mode` flag for troubleshooting and a `/cd` command for changing directories mid-session without disrupting the prompt cache. It also includes numerous bug fixes addressing issues with UI responsiveness, enterprise policies, and agent management across platforms. These changes benefit developers and architects troubleshooting complex setups or working with large codebases, offering a more stable and efficient experience.
patch feature - Anthropic Python SDK Releases sdkaiengineer ·
Anthropic SDK Python v0.107.1 fixes API key header issue
This release of the Anthropic SDK for Python, version 0.107.1, addresses a bug related to API key authentication. Specifically, it ensures the 'x-api-key' header is correctly sent when using API key authentication with Foundry. This fix is important for users experiencing authentication failures due to the missing header.
patch - Claude Code Releases sdkaiengineer ·
Claude-Code v2.1.168: Bug fixes and reliability improvements
Claude-Code version 2.1.168 has been released, focusing on bug fixes and general reliability enhancements. This patch aims to improve the stability and performance of the Claude-Code model for its users. The update is available now for those who use Claude-Code in their development workflows.
patch - Claude Code Releases sdkaiengineer ·
Claude-Code v2.1.167: Bug Fixes and Reliability Improvements
Claude-Code version 2.1.167 addresses underlying bugs and enhances overall reliability. This patch is relevant for users of the Claude-Code model, particularly those integrating it into their development workflows. The specific improvements focus on fixing defects rather than introducing new features.
patch - Claude Code Releases sdkaiengineermedia ·
Claude Code v2.1.166: Improved fallbacks, session hardening, and bug fixes
This patch for Claude Code introduces a `fallbackModel` setting to automatically retry with alternative models on overload or unavailability, enhancing resilience. It also hardens cross-session messaging by stripping user authority from relayed messages and adds improved error handling for API rejections. These changes benefit users relying on Claude Code for coding assistance, particularly in complex or unstable environments, by ensuring more consistent operation and fixing several bugs affecting IDE integration and terminal handling.
patch - Anthropic Python SDK Releases sdkaideprecation ·
Anthropic SDK Python v0.106.0: Claude Opus 4.1 deprecated, client fixes
Anthropic has released version 0.106.0 of its Python SDK, marking Claude Opus 4.1 as deprecated. This release also includes bug fixes for the Foundry client's copy and with_options methods, and improves schema transformation to preserve $defs. Users of the Python SDK are advised to note the deprecation and benefit from the client enhancements.
feature patch - Claude Code Releases sdkaiengineer ·
Claude Code v2.1.165 Patch Addresses Bugs
Anthropic has released version 2.1.165 of Claude Code, focusing on bug fixes and reliability improvements. This patch aims to enhance the stability and performance of the coding assistant for users. It is available now for all users leveraging Claude Code.
patch - Claude Platform Release Notes aideprecation ·
Claude Platform Updates: New Models, Deprecations, and Tool Enhancements
Anthropic has released new flagship models, Claude Fable 5 and Claude Opus 4.8, featuring larger context windows and enhanced adaptive thinking. The company is also retiring older models like Claude Sonnet 4 and Opus 4.1, recommending upgrades. Several tools and platform features, including code execution, web search, and managed agents, have received updates and expanded availability on AWS, impacting developers and researchers using the Claude API.
feature deprecation patch - Claude Code Releases sdkaiengineer ·
Claude Code v2.1.162: Enhanced Agent Features and Bug Fixes
This update for Claude Code introduces several enhancements, including new details in `claude agents --json` to clarify session blocking, improved handling of search tools, and persistent default settings for effort levels. It also addresses numerous bugs, such as silent startup hangs, incorrect permission rule application, and display issues within the agents view, improving overall stability and user feedback for developers using the integrated coding environment.
patch feature - Anthropic News ·
AI-enabled cyber threats analysis by Anthropic's Frontier Red Team
Anthropic's Frontier Red Team analyzed a year's worth of AI-enabled cyber threats, mapping attacker techniques to the MITRE ATT&CK framework. The analysis reveals that malicious actors are increasingly using AI for complex, autonomous operations, challenging traditional risk assessment methods. These findings are informing the development of safeguards for Anthropic's models and discussions with MITRE to evolve the ATT&CK framework.
announcement - Anthropic News ai ·
Anthropic Launches Claude Partner Network Services Track and Partner Hub
Anthropic is expanding its Claude Partner Network by introducing a Services Track with tiered levels and a Partner Hub portal. These initiatives aim to provide customers with clearer visibility into partner capabilities and qualifications for deploying AI in production. The program requires partners to meet specific criteria for certified individuals, deployed customers, and public endorsements, ensuring a consistent standard for firms of all sizes.
announcement feature - Claude Code Releases sdkai ·
Claude Code v2.1.161: Enhanced Usage Metrics, Tool Call Fixes, and UI Improvements
Claude Code version 2.1.161 incorporates `OTEL_RESOURCE_ATTRIBUTES` into metric datapoints for granular usage analysis, and enhances parallel tool call reliability by preventing failed Bash commands from canceling other operations. This patch also introduces UI refinements, improves handling of unused connectors, and addresses several regressions and bugs affecting sessions, policies, and terminal rendering for a smoother developer experience. The update is available now for all users.
patch feature - Claude Code Releases sdkai ·
Claude Code v2.1.160: Security, Stability, and UX Improvements
This patch enhances Claude Code's security and stability by adding prompts for writing to sensitive shell startup and build-tool config files, preventing unintended command execution. It also improves user experience with fixes for session restoration, clipboard functionality on WSL, and responsiveness issues on Windows. These updates benefit all users, particularly those on Windows or relying on background sessions, with improved stability and a more robust security posture.
patch security - Claude Platform Release Notes aimedia ·
Claude Platform Enhancements: Output Capping and Refusal Billing
Claude Platform now supports a `max_tokens` parameter for the advisor tool to reduce latency and costs by capping output. Additionally, users are no longer billed for requests that result in a refusal without any output. These changes are available on the Claude API and aim to improve efficiency and cost management for developers utilizing the advisor tool and handling refusals.
feature patch - Anthropic News aisecurityengineer ·
Anthropic Expands Project Glasswing to 150 New Organizations
Anthropic is expanding Project Glasswing, an initiative to secure critical software, to approximately 150 new organizations across over 15 countries. This expansion aims to leverage AI, specifically Claude Mythos Preview, to scan codebases for vulnerabilities and promote more secure software development practices in the face of increasingly powerful AI models. The program now includes organizations in diverse sectors like power, water, and healthcare, and emphasizes the need for robust safeguards as AI capabilities advance. New partners will undergo security reviews before gaining access.
announcement feature