https://martin-minghetti.github.io/modelsentry Automated AI news and alerts for developers https://ai.google.dev/pricing https://ai.google.dev/pricing Wed, 22 Apr 2026 09:09:02 GMT A diff alert for Google Gemini's pricing page reveals an update noting the availability of 'Gemini Deep Research' in preview, along with details on new features like collaborative planning and MCP support. | Impact: medium | Signal: positive | Why included: This diff alert indicates a release of new features and preview availability for Google's Gemini API, signifying product development. novedad https://www.anthropic.com/pricing https://www.anthropic.com/pricing Wed, 22 Apr 2026 09:09:00 GMT A diff alert for Anthropic's pricing page shows no significant changes detected, indicating stability in their current pricing structure. | Impact: low | Signal: neutral | Why included: This diff alert shows no change in Anthropic's pricing, which is relevant for monitoring for potential future changes. novedad https://www.theguardian.com/technology/2026/apr/22/anthropic-investigates-report-of-rogue-access-to-hack-enabling-mythos-ai https://www.theguardian.com/technology/2026/apr/22/anthropic-investigates-report-of-rogue-access-to-hack-enabling-mythos-ai Wed, 22 Apr 2026 09:06:40 GMT Anthropic is looking into reports that unauthorized users gained access to its Mythos AI model, which is described as 'hack-enabling'. This incident raises security concerns regarding the accessibility and control of advanced AI models. | Impact: high | Signal: negative | Why included: This is a high-impact security alert involving unauthorized access to a potentially sensitive AI model. alerta https://news.ycombinator.com/item?id=47860903 https://news.ycombinator.com/item?id=47860903 Wed, 22 Apr 2026 08:53:53 GMT A user is seeking open-source alternatives for Claude's design, having found only one GitHub project so far. This indicates community interest in exploring different design approaches for AI systems. | Impact: low | Signal: neutral | Why included: This is a community discussion about open-source design alternatives for AI models. novedad https://www.gsarigiannidis.gr/claude-global-memory/ https://www.gsarigiannidis.gr/claude-global-memory/ Wed, 22 Apr 2026 07:36:39 GMT This article discusses globalizing Claude's project memories, suggesting an exploration of how AI models manage and potentially share project-related information across time and contexts. | Impact: low | Signal: neutral | Why included: This is a discussion about AI model memory and project management, relevant to the LLM industry's ongoing development. novedad https://sunglasses.dev/reports/anthropic-cvp-opus-4-7-evaluation-run-2 https://sunglasses.dev/reports/anthropic-cvp-opus-4-7-evaluation-run-2 Wed, 22 Apr 2026 07:30:20 GMT This is a report on evaluation run 2 for Anthropic's CVP, likely referring to a model or system performance evaluation, indicating ongoing model development and testing by Anthropic. | Impact: low | Signal: neutral | Why included: This is a technical report on Anthropic's model evaluation, relevant to understanding their development progress. novedad https://claude.com/docs/cowork/3p/overview https://claude.com/docs/cowork/3p/overview Wed, 22 Apr 2026 06:32:16 GMT This documentation provides an overview of Claude Cowork, detailing how to use it with your own cloud inference provider, which is significant for enterprise adoption and flexibility. | Impact: low | Signal: neutral | Why included: This is documentation for a feature that enables flexible deployment and integration of Claude's capabilities. novedad https://github.com/m-ghalib/gemini-plugin-cc https://github.com/m-ghalib/gemini-plugin-cc Wed, 22 Apr 2026 04:41:09 GMT A developer created a Gemini plugin for Claude Code, allowing Claude Code to delegate tasks to the Gemini CLI. This demonstrates cross-platform integration and leveraging complementary AI capabilities. | Impact: medium | Signal: positive | Why included: This is a significant feature launch demonstrating integration between major LLM providers, expanding functionality. novedad https://raxitlabs.com/blogs/claude-47-five-layers-cyber-blocking https://raxitlabs.com/blogs/claude-47-five-layers-cyber-blocking Wed, 22 Apr 2026 03:31:00 GMT This article discusses how Claude 4.7 handles cybersecurity prompts, detailing its proactive and reactive blocking mechanisms. It sheds light on the security features and safety measures implemented in advanced LLMs. | Impact: low | Signal: neutral | Why included: This is a discussion about the security and safety features of a specific LLM version. novedad https://github.com/Screenata/open-chronicle https://github.com/Screenata/open-chronicle Wed, 22 Apr 2026 03:06:15 GMT A developer built an open-source version of OpenAI Chronicle, focusing on local screen memory for Claude Code and Codex CLI. This highlights advancements in local AI tools and user privacy. | Impact: medium | Signal: positive | Why included: This is a launch of a new open-source tool for local AI, indicating growth in the privacy-focused AI tooling sector. novedad https://simonwillison.net/2026/Apr/22/claude-code-confusion/ https://simonwillison.net/2026/Apr/22/claude-code-confusion/ Wed, 22 Apr 2026 02:52:56 GMT This article discusses confusion surrounding the pricing of Claude Code, with speculation about a $100/month cost. It highlights the complexity and potential lack of clarity in AI service pricing models. | Impact: medium | Signal: negative | Why included: This alerts to potential pricing changes or confusion for a significant AI product, impacting user costs. alerta https://news.ycombinator.com/item?id=47857883 https://news.ycombinator.com/item?id=47857883 Wed, 22 Apr 2026 02:07:12 GMT A new SDK, govern(), has been released by AgentLattice to govern Anthropic Managed Agents with just three lines of code. This offers enhanced control and security for enterprise AI deployments. | Impact: medium | Signal: positive | Why included: This is a release of a new tool that enhances governance and security for enterprise AI deployments. novedad https://www.wsj.com/us-news/law/openai-under-criminal-probe-in-florida-over-mass-shooters-chatgpt-use-47913814 https://www.wsj.com/us-news/law/openai-under-criminal-probe-in-florida-over-mass-shooters-chatgpt-use-47913814 Wed, 22 Apr 2026 00:49:01 GMT OpenAI is reportedly under criminal investigation in Florida due to the use of ChatGPT by a mass shooter. This raises serious ethical and legal questions regarding AI misuse and accountability. | Impact: high | Signal: negative | Why included: This is a high-impact legal and ethical alert concerning AI misuse and its potential societal consequences. alerta https://www.theregister.com/2026/04/22/anthropic_removes_claude_code_pro/ https://www.theregister.com/2026/04/22/anthropic_removes_claude_code_pro/ Wed, 22 Apr 2026 00:44:39 GMT Anthropic is reportedly testing developer reactions to removing Claude Code from its Pro plan. This could significantly impact user access and pricing for AI coding tools. | Impact: high | Signal: negative | Why included: This is a high-impact alert as it suggests a potential removal of a feature from a popular AI product, impacting users. alerta https://www.reuters.com/technology/anthropics-mythos-model-accessed-by-unauthorized-users-bloomberg-news-reports-2026-04-21/ https://www.reuters.com/technology/anthropics-mythos-model-accessed-by-unauthorized-users-bloomberg-news-reports-2026-04-21/ Wed, 22 Apr 2026 00:40:07 GMT Bloomberg reports that unauthorized users have gained access to Anthropic's Mythos model, raising significant security concerns about the containment and access controls of advanced AI models. | Impact: high | Signal: negative | Why included: This is a high-impact security alert regarding unauthorized access to a specific AI model, indicating potential vulnerabilities. alerta https://news.ycombinator.com/item?id=47856900 https://news.ycombinator.com/item?id=47856900 Wed, 22 Apr 2026 00:24:17 GMT Ravix is a new AI agent that operates on existing Claude Code subscriptions, offering an autonomous agent experience without additional API keys or billing. This simplifies agent deployment for users. | Impact: medium | Signal: positive | Why included: This is a release of a new AI agent that leverages existing subscriptions, indicating innovation in AI agent accessibility. novedad https://winners.webbyawards.com/2026/specialachievement/476/claude https://winners.webbyawards.com/2026/specialachievement/476/claude Wed, 22 Apr 2026 00:19:06 GMT Claude has been recognized as the Webby Person of the Year, a significant achievement highlighting its impact and recognition in the digital and AI space. | Impact: medium | Signal: positive | Why included: This is a significant award recognizing the impact and influence of Claude in the AI industry. novedad https://techcrunch.com/2026/04/21/unauthorized-group-has-gained-access-to-anthropics-exclusive-cyber-tool-mythos-report-claims/ https://techcrunch.com/2026/04/21/unauthorized-group-has-gained-access-to-anthropics-exclusive-cyber-tool-mythos-report-claims/ Wed, 22 Apr 2026 00:15:41 GMT Multiple reports indicate that an unauthorized group accessed Anthropic's exclusive cyber tool, Mythos. This raises serious security concerns and highlights potential vulnerabilities in sensitive AI tools. | Impact: high | Signal: negative | Why included: This is a high-impact security alert regarding unauthorized access to a specialized AI tool, indicating potential risks. alerta https://www.startuphub.ai/agents/submit https://www.startuphub.ai/agents/submit Wed, 22 Apr 2026 00:13:49 GMT A tool is available that uses Claude or ChatGPT to help submit startup profiles in under 60 seconds. This showcases the application of LLMs in streamlining business processes. | Impact: low | Signal: positive | Why included: This demonstrates a practical application of LLMs in automating business processes, relevant to AI adoption. novedad https://news.ycombinator.com/item?id=47856073 https://news.ycombinator.com/item?id=47856073 Tue, 21 Apr 2026 23:14:09 GMT Claude Code has reportedly been removed from the Pro plan, a change that could impact users who relied on this feature within their subscription. | Impact: high | Signal: negative | Why included: This is a high-impact alert as it indicates a feature removal from a paid plan, potentially affecting user access and value. alerta https://www.wheresyoured.at/news-anthropic-removes-pro-cc/ https://www.wheresyoured.at/news-anthropic-removes-pro-cc/ Tue, 21 Apr 2026 22:56:09 GMT Claude Code has been removed from the $20/month 'Pro' subscription for new users. This change signifies a potential shift in Anthropic's product strategy and pricing. | Impact: high | Signal: negative | Why included: This is a high-impact alert concerning a pricing change that affects feature availability for a significant user base. alerta https://bsky.app/profile/edzitron.com/post/3mjzxwfx3qs2a https://bsky.app/profile/edzitron.com/post/3mjzxwfx3qs2a Tue, 21 Apr 2026 22:36:24 GMT There are discussions and reports suggesting that Claude Code might be removed from the Pro tier subscription. This potential change is causing significant community discussion. | Impact: high | Signal: negative | Why included: This alert signals a potential feature removal from a paid tier, which could significantly impact user perception and value. alerta https://openai.com/index/scaling-codex-to-enterprises-worldwide https://openai.com/index/scaling-codex-to-enterprises-worldwide Tue, 21 Apr 2026 00:00:00 GMT OpenAI has launched Codex Labs and partnered with major consulting firms to facilitate enterprise adoption of Codex. The product has also reached 4 million weekly active users, indicating significant market traction. | Impact: medium | Signal: positive | Why included: This indicates a significant partnership and milestone for OpenAI's enterprise offerings, signaling strong market adoption. novedad https://github.com/anthropics/claude-code/releases/tag/v2.1.116 https://github.com/anthropics/claude-code/releases/tag/v2.1.116 Mon, 20 Apr 2026 22:18:28 GMT Claude Code version 2.1.116 introduces significant performance improvements for '/resume' on large sessions, faster MCP startup, smoother fullscreen scrolling in VS Code, Cursor, and Windsurf terminals, and an improved thinking spinner for inline progress. It also includes security enhancements, bug fixes for rendering and keyboard shortcuts, and improvements to slash command behavior and plugin dependency management. | Impact: low | Signal: positive | Why included: This is a significant update to a developer tool for a major LLM, introducing performance and security improvements. novedad https://simonwillison.net/2026/Apr/20/llm-openrouter/ https://simonwillison.net/2026/Apr/20/llm-openrouter/ Mon, 20 Apr 2026 18:00:26 GMT An update to the llm-openrouter library to version 0.6 has been released. This likely includes improvements or changes to how LLMs are accessed via OpenRouter. | Impact: low | Signal: neutral | Why included: This is a library update relevant to accessing LLMs through a specific platform. novedad https://beyondmachines.net/event_details/anthropic-claude-code-leak-reveals-critical-command-injection-vulnerabilities-e-6-c-1-k/gD2P6Ple2L https://beyondmachines.net/event_details/anthropic-claude-code-leak-reveals-critical-command-injection-vulnerabilities-e-6-c-1-k/gD2P6Ple2L Mon, 20 Apr 2026 07:20:38 GMT A leak of Anthropic's Claude Code has exposed critical command injection vulnerabilities within the system. | Impact: high | Signal: negative | Why included: A security vulnerability in a major LLM model is a high-impact event. alerta https://github.com/Oaklight/llm-rosetta https://github.com/Oaklight/llm-rosetta Mon, 20 Apr 2026 07:05:33 GMT A new tool, LLM-Rosetta, has been developed to translate LLM API calls between different providers like OpenAI, Anthropic, and Gemini, aiming to simplify adapter maintenance. | Impact: medium | Signal: positive | Why included: A new tool simplifying API integration across major LLM providers is a notable development in the ecosystem. novedad https://www.wired.com/story/openai-executive-kevin-weil-is-leaving-the-company/ https://www.wired.com/story/openai-executive-kevin-weil-is-leaving-the-company/ Mon, 20 Apr 2026 07:01:56 GMT OpenAI executive Kevin Weil is departing the company. | Impact: low | Signal: neutral | Why included: Executive departures at major AI companies can sometimes signal shifts in strategy or focus, though this is a low-impact event on its own. novedad https://simianwords.bearblog.dev/conclusive-proofs-that-llm-costs-are-going-down/ https://simianwords.bearblog.dev/conclusive-proofs-that-llm-costs-are-going-down/ Mon, 20 Apr 2026 05:30:00 GMT An article argues that LLM costs are actually decreasing, contrary to potential assumptions. | Impact: medium | Signal: positive | Why included: Information about LLM cost trends is crucial for understanding market dynamics and adoption rates. novedad https://www.axios.com/2026/04/19/nsa-anthropic-mythos-pentagon https://www.axios.com/2026/04/19/nsa-anthropic-mythos-pentagon Mon, 20 Apr 2026 05:05:13 GMT The NSA is reportedly using Anthropic's Mythos, despite the model being blacklisted. | Impact: medium | Signal: negative | Why included: The use of a blacklisted model by a government agency raises security and compliance concerns. alerta https://www.thatprivacyguy.com/blog/anthropic-spyware/ https://www.thatprivacyguy.com/blog/anthropic-spyware/ Mon, 20 Apr 2026 02:41:50 GMT A user alleges that Anthropic installed a 'spyware bridge' on their machine. | Impact: high | Signal: negative | Why included: Allegations of spyware from a major AI company are a serious security concern. alerta https://simonwillison.net/2026/Apr/20/claude-token-counts/ https://simonwillison.net/2026/Apr/20/claude-token-counts/ Mon, 20 Apr 2026 00:56:40 GMT A Claude token counter tool has been updated to include comparisons across different models. | Impact: medium | Signal: positive | Why included: A tool that facilitates model comparison for token usage is useful for developers optimizing costs and performance. novedad https://simonwillison.net/2026/Apr/20/claude-token-counts/#atom-everything https://simonwillison.net/2026/Apr/20/claude-token-counts/#atom-everything Mon, 20 Apr 2026 00:50:45 GMT The Claude Token Counter tool has been updated to include comparisons between different Claude models. | Impact: low | Signal: positive | Why included: This is a useful tool update for developers working with Claude, providing added functionality for model comparison. novedad https://openai.com/index/hyatt-advances-ai-with-chatgpt-enterprise https://openai.com/index/hyatt-advances-ai-with-chatgpt-enterprise Mon, 20 Apr 2026 00:00:00 GMT Hyatt is deploying ChatGPT Enterprise across its global workforce, leveraging GPT-5.4 and Codex to enhance productivity, streamline operations, and improve guest experiences. | Impact: medium | Signal: positive | Why included: This signifies a major enterprise adoption of OpenAI's technology, indicating real-world business value and market penetration. novedad https://bannedbyanthropic.com/ https://bannedbyanthropic.com/ Sun, 19 Apr 2026 22:33:31 GMT A website appears to list users who have been banned by Anthropic. | Impact: low | Signal: neutral | Why included: Information about platform bans can be relevant for users interacting with AI services. novedad https://news.ycombinator.com/item?id=47828154 https://news.ycombinator.com/item?id=47828154 Sun, 19 Apr 2026 22:20:17 GMT A user suggests adding tagging functionality to the Claude UI to improve conversation searching and organization. | Impact: low | Signal: positive | Why included: User feedback on UI improvements for major LLMs is relevant for product development and user experience. novedad https://tidbits.com/2026/04/09/what-anthropics-mythos-and-project-glasswing-mean-for-your-apple-devices/ https://tidbits.com/2026/04/09/what-anthropics-mythos-and-project-glasswing-mean-for-your-apple-devices/ Sun, 19 Apr 2026 20:09:03 GMT An analysis explores the implications of Anthropic's Mythos and Project Glasswing for Apple devices. | Impact: medium | Signal: neutral | Why included: New developments from Anthropic impacting major hardware platforms like Apple are of significant interest. novedad https://codeberg.org/mlow/lmcli https://codeberg.org/mlow/lmcli Sun, 19 Apr 2026 19:58:05 GMT Lmcli v0.5.0 has been released, described as a usable tool for LLM minimalists, now featuring image support. | Impact: medium | Signal: positive | Why included: A new version of an LLM tool with added image support is a relevant update for developers working with LLMs. novedad https://wecreatethis.com/blog/post?slug=claude-is-skeptical-about-openclaw https://wecreatethis.com/blog/post?slug=claude-is-skeptical-about-openclaw Sun, 19 Apr 2026 08:25:24 GMT This blog post discusses Claude's perceived skepticism towards a technology or project named OpenClaw. | Impact: low | Signal: neutral | Why included: This item is relevant as it mentions Claude, a prominent LLM, in the context of a discussion about another technology. novedad https://github.com/JoJa84/Codefone https://github.com/JoJa84/Codefone Sun, 19 Apr 2026 07:10:54 GMT Android 15 includes a hidden Linux Terminal that functions as a Debian VM and can run Claude Code. | Impact: medium | Signal: positive | Why included: This is relevant as it showcases the integration of LLM code (Claude Code) within a mobile operating system's terminal, indicating broader application and accessibility. novedad https://github.com/adityaarakeri/fuelgauge https://github.com/adityaarakeri/fuelgauge Sun, 19 Apr 2026 06:07:04 GMT A new tool called Fuelgauge provides a Claude Code status line that operates without requiring Node.js. | Impact: medium | Signal: positive | Why included: This is relevant as it's a new tool for Claude Code, indicating ecosystem development and improved developer experience. novedad https://agentrq.com https://agentrq.com Sun, 19 Apr 2026 02:44:18 GMT AgentRQ is a real-time platform for agent-human task collaboration, supporting Claude Code and Gemini. | Impact: medium | Signal: positive | Why included: This is relevant as it demonstrates advancements in multi-agent systems and collaboration between different LLMs, specifically Claude and Gemini. novedad https://www.artificialintelligencemadesimple.com/p/chatgpt-vs-gemini-vs-claude-the-best https://www.artificialintelligencemadesimple.com/p/chatgpt-vs-gemini-vs-claude-the-best Sun, 19 Apr 2026 02:38:48 GMT An article comparing ChatGPT, Gemini, and Claude to determine the best LLM subscription. | Impact: low | Signal: neutral | Why included: This item is relevant for comparing major LLM providers, offering insights into market positioning and user choice. novedad https://github.com/noahkirsch/Claude-Code-Glow https://github.com/noahkirsch/Claude-Code-Glow Sun, 19 Apr 2026 02:38:44 GMT A user shares personal methods for improving Claude Code's efficiency by at least 30%. | Impact: medium | Signal: positive | Why included: This is relevant as it provides practical insights into optimizing LLM performance and efficiency, directly impacting user experience and resource utilization. novedad https://github.com/salaamalykum/quran-semantic-search https://github.com/salaamalykum/quran-semantic-search Sun, 19 Apr 2026 02:07:34 GMT A 5-translation RAG matrix is presented as a solution to fix religious hallucinations in LLMs. | Impact: medium | Signal: positive | Why included: This item addresses a critical issue of LLM hallucinations, particularly in sensitive areas like religious content, and proposes a technical solution. novedad https://www.reddit.com/r/ClaudeAI/comments/1spg4s9/ai_code_base_is_so_fragile/ https://www.reddit.com/r/ClaudeAI/comments/1spg4s9/ai_code_base_is_so_fragile/ Sun, 19 Apr 2026 02:07:05 GMT A user expresses disappointment with the fragility and unreliability of AI-generated codebases for production-level applications. | Impact: medium | Signal: negative | Why included: This is relevant as it raises concerns about the practical usability of AI-generated code for complex, real-world applications, highlighting potential limitations in reliability. novedad https://github.com/NVIDIA/TensorRT-LLM https://github.com/NVIDIA/TensorRT-LLM Sun, 19 Apr 2026 01:58:00 GMT A GitHub repository for TensorRT LLM, likely related to NVIDIA's optimizations for LLM inference. | Impact: low | Signal: neutral | Why included: This is relevant as TensorRT is a key technology for optimizing deep learning model performance, and its application to LLMs is significant for efficiency. novedad https://techcrunch.com/2026/04/12/at-the-humanx-conference-everyone-was-talking-about-claude/ https://techcrunch.com/2026/04/12/at-the-humanx-conference-everyone-was-talking-about-claude/ Sun, 19 Apr 2026 00:52:49 GMT An article highlighting the significant presence and discussion of Claude at the HumanX conference. | Impact: low | Signal: positive | Why included: This shows Claude's growing influence and mindshare within the AI community and at industry events. novedad https://github.com/yasyf/slop-cop https://github.com/yasyf/slop-cop Sun, 19 Apr 2026 00:20:47 GMT A Cursor plugin called Slop Cop for Claude Code is presented. | Impact: medium | Signal: positive | Why included: This is relevant as it's a new plugin for Claude Code, indicating ecosystem development and new tools for developers using Claude. novedad https://simonwillison.net/2026/Apr/18/opus-system-prompt/ https://simonwillison.net/2026/Apr/18/opus-system-prompt/ Sun, 19 Apr 2026 00:14:06 GMT An analysis of the changes made to the system prompt between Claude Opus versions 4.6 and 4.7. | Impact: medium | Signal: neutral | Why included: This is relevant as it details changes in system prompts for a specific LLM version, offering insights into how LLMs are being tuned and updated. novedad