Radar
Weekly signals across AI research, releases, and infrastructure. Filtered for what matters when you're building with it.
Pulse
Weekly indicators across AI research and open-weight releases
arXiv Papers →
cs.AI + cs.CL + cs.LG
892-5% vs 4w avg
Open-Weight Releases →
Notable models per week
4-6% vs 4w avg
Shifts strategyWorth watching↑↓Theme momentum vs 4-week avg
DateSourceTitle
OpenAIThe Next Phase of Enterprise AIAI NewsWhy Companies Like Apple Are Building AI Agents With LimitsAnthropic ResearchTrustworthy Agents in PracticeAI NewsAgentic AI's Governance Challenges Under the EU AI Act in 2026BloombergUS Officials Summon Bank CEOs Over Anthropic Mythos CapabilitiesAnthropicAnthropic Launches Claude Managed Agents in Public BetaMeta AIMeta Launches Muse Spark With Multi-Agent Orchestration, Drops Open SourceAnthropicAnthropic Releases Claude Mythos Preview via Project GlasswingAnthropicAnthropic Expands to 1M Google TPUs, Passes $30B Revenue Run RateDateSourceTitle
Meta AIMeta Releases Llama 4 Scout and Maverick: Open MoE With 10M ContextOpenAIOpenAI Begins Sora Shutdown: App Closes April 26, API in SeptemberAnthropic / TNWAnthropic Blocks Third-Party Frameworks From Claude SubscriptionsGoogle DeepMindGoogle Releases Gemma 4: Open Trimodal Models With 256K ContextOpenAIOpenAI Codex Adds Pay-as-You-Go Pricing for TeamsAnthropic APIAnthropic Raises Message Batches Output Cap to 300K TokensOpenAIOpenAI Closes $122B Round at $852B ValuationDateSourceTitle
Fortune / AnthropicAnthropic Confirms 'Claude Mythos' After Accidental Data LeakMistral AIMistral Releases Voxtral TTS: Open-Weight 4B Streaming Speech ModelAnthropic ResearchAnthropic Economic Index: Learning CurvesOpenAIOpenAI Expands Agentic Commerce: Walmart In-ChatGPT Shopping Goes LiveBloomberg / OpenAIOpenAI Discontinues Sora App and APIarXivFinMCP-Bench: Benchmarking LLM Agents for Real-World Financial Tool Use Under MCPAI NewsAutomating Complex Finance Workflows With Multimodal AIDateSourceTitle
NVIDIA ResearchNemotron-Cascade 2: Open 30B MoE with 3B Active ParametersOpenAIIntroducing GPT-5.4 mini and nanoMistral AINVIDIA Launches Nemotron Coalition with Mistral, Perplexity, and 6 More LabsarXivClawWorm: Self-Propagating Attacks Across LLM Agent EcosystemsarXivSAGE: Multi-Agent Self-Evolution for LLM ReasoningarXivSpend Less, Reason Better: Budget-Aware Value Tree Search for LLM AgentsDateSourceTitle
HuggingFace PapersDIVE: Scaling Diversity in Agentic Task Synthesis for Generalizable Tool UseAnthropic1M Token Context Window Now Generally Available for Claude Opus 4.6 and Sonnet 4.6AnthropicAnthropic Invests $100 Million into the Claude Partner NetworkAI NewsHow Multi-Agent AI Economics Influence Business AutomationOpenAIGoogle and Microsoft Apps in ChatGPT Now Support Write ActionsMistral AIRails Testing on Autopilot: Building an Agent That Writes What Developers Won'tAI NewsManulife Moves AI Agents into Core Financial WorkflowsAI NewsMastercard Completes First Live Authenticated Agentic Transaction with DBS and UOBGitHubPageAgent: In-Page GUI Agent That Controls Web Interfaces with Natural LanguageGitHubagency-agents: Complete AI Agency with Specialized AgentsGitHubHermes Agent: The Agent That Grows With YouDateSourceTitle
arXivMASFactory: Graph-Centric Framework for Multi-Agent Orchestration with Vibe GraphingOpenAIReasoning Models Struggle to Control Their Chains of ThoughtOpenAICodex Security: Application Security Agent in Research PreviewOpenAISkills for ChatGPT Business & Enterprise: Reusable Workflow AutomationOpenAIGPT-5.4 Thinking & Pro: Unified Reasoning + Coding + Agentic ModelAnthropic ResearchLabor Market Impacts of AI: A New Measure and Early EvidenceAI NewsDyna.Ai Raises Series A for Agentic AI in Financial ServicesAI NewsScaling Intelligent Automation Without Breaking Live WorkflowsDateSourceTitle
OpenAIOpenAI Raises $110B at $730B ValuationTechCrunchAnthropic Launches Enterprise Agent Plugins for Finance, Engineering, and DesignOpenAIOpenAI Introduces Lockdown Mode and Elevated Risk Labels in ChatGPTarXivARLArena: A Unified Framework for Stable Agentic Reinforcement LearningTechCrunchGoogle Labs Launches Agent Step in Opal to Build Agentic AI WorkflowsHuggingFace PapersRAG-Anything: Unified Multimodal Knowledge RetrievalDateSourceTitle
Anthropic2026 Agentic Coding Trends Report