Models & Labs
Claude Fable 5 now available in GitHub Copilot
GitHub Changeloghigh confidence
Why it matters
- →Claude Fable 5 introduces a new approach to data retention for safety in AI models.
- →It enhances coding efficiency with fewer tool calls and lower token consumption.
- →This model expands the capabilities of GitHub Copilot for enterprise users.
Anthropic's Claude Fable 5 model is now available in GitHub Copilot, marking the first Mythos class model designed for autonomous coding and knowledge tasks. Unlike previous models, Claude Fable 5 requires data retention for up to 30 days to operate safety classifiers, a departure from the zero data retention policy of other Claude models. This model is available to Copilot Pro+, Max, Business, and Enterprise users and promises more efficient coding with fewer tool calls. The rollout will be gradual, and administrators must enable the policy to access the model.
Read originalMore from GitHub Changelog
© GitHub ChangelogCoding Toolscoding
GitHub CLI v2.94.0 Enhances Issue Management
GitHub CLI's latest update, v2.94.0, significantly enhances issue management by allowing developers to handle issue types, sub-issues, and dependencies directly from the terminal. This means developers can now set issue types, manage parent and sub-issue relationships, and mark dependencies without leaving the command line. The update also introduces new JSON fields in 'gh issue view' and 'gh issue list', enabling automation scripts to interact with issue structures more effectively. This streamlines workflows for developers and coding agents, reducing the need to switch between the CLI and browser.
GitHub Changelog
© GitHub ChangelogCoding Toolscoding
GitHub Copilot Chat Enhances Agent Session Management
GitHub has enhanced its Copilot Chat by integrating better session management with the Copilot cloud agent. Users can now search and query past agent sessions directly within the chat, making it easier to track and continue work. The update introduces tools like 'Get agent logs' for reviewing session changes and 'Session search' for finding past sessions by topic or recency. This improvement streamlines the workflow for developers, allowing them to manage and follow up on their coding tasks more efficiently.
GitHub Changelog
© GitHub ChangelogCoding Toolscoding
GitHub Copilot CLI Adds Security Review Command
GitHub has introduced a new security review command in its Copilot CLI, offering developers an AI-driven tool to identify security vulnerabilities in their code changes. This experimental feature, available in public preview, allows users to run a security scan directly from the terminal, providing high-confidence findings and actionable suggestions. Unlike GitHub's existing security tools, this command offers a lightweight, on-demand review process that integrates seamlessly into developers' workflows. This addition enhances the security toolkit available to developers, enabling them to catch potential issues before code reaches production.
GitHub Changelog
More in Models & Labs
© The Verge AIModels & Labsmodels
Anthropic's Claude Fable 5 Limits Biology Queries
Anthropic's release of Claude Fable 5, touted as their most powerful AI model, comes with significant limitations in answering biology-related questions. This is due to the model's conservative safeguards designed to prevent misuse in bioweapons research. While the model excels in cybersecurity tasks, its biology filters are so stringent that even basic queries like 'what are mitochondria' are blocked. Anthropic aims to balance safety with utility, promising future adjustments to reduce false positives and potentially open up more capabilities for scientific research.
The Verge AI
© Google DeepMindModels & Labsmodels
Google DeepMind unveils DiffusionGemma for faster text generation
Google DeepMind's DiffusionGemma marks a significant shift in text generation by leveraging diffusion techniques to generate text blocks up to four times faster than traditional models. This 26B Mixture of Experts model, designed for speed-critical applications, moves beyond the sequential token-by-token approach, allowing for parallel generation of 256 tokens. While it offers blazing fast inference on GPUs, it trades off some quality compared to the standard Gemma 4 models. This innovation is particularly beneficial for developers working on real-time interactive AI applications, as it maximizes hardware utilization and reduces latency bottlenecks.
Google DeepMind
© Hugging Face BlogModels & Labsmodels
Cohere Launches North Mini Code Model for Developers
Cohere has unveiled North Mini Code, a 30B-parameter Mixture-of-Experts model designed for complex software engineering tasks, now available on Hugging Face. This model stands out with its agentic coding capabilities, optimized for terminal-based tasks and high-quality code generation. It outperforms several larger models in coding benchmarks, showcasing its efficiency and robustness. By employing a unique training approach with supervised fine-tuning and reinforcement learning, North Mini Code aims to serve as a reliable foundation for coding agents. This release marks a significant step in making advanced coding models accessible to developers.
Hugging Face Blog