📂 Category
Research & Benchmarks Articles - Complete AI News Archive
932 articles in this category • Page 1 of 10
- 1. Researchers Plant 2 Million Digital Tripwires in AI Services
- 2. Clairic AI Reaches 98% Accuracy on Customer Support, Receipt Data
- 3. Announcing Data Annotation Queues: LLM Apps Can Start Without Datasets
- 4. LangSmith Tool Gateway Manages Access, Identity and Policies for Every Tool Call
- 5. AI Scans Thousands of Medical Papers for Shared Code Repositories
- 6. Meta's Continual Learning Agent Hits 56.0 Score With Prompt-Swapping Technique
- 7. OpenAI agent built unauthorized email surveillance tool during writing task
- 8. Google's Gemini 4 Neon AI Model Has One-Million-Token Context
- 9. OpenAI researchers value their AI tools more than half their colleagues
- 10. Gemini 4 Neon Review: How It Ranks on Key AI Benchmarks
- 11. Sumitomo Life Leaders Ease ChatGPT Enterprise Adoption for Teams
- 12. Auditors Struggle to Catch Sabotage in Full AI Safety Research Transcripts
- 13. OpenAI says its models boosted AMD chip design, accelerated custom chip development
- 14. Google Deepmind's Genie 4 update adds chainable worlds, remix gallery
- 15. AI Models Flagged Medics, Journalists as Drone Targets
- 16. The Download: AI's New Sense of Smell and Hydropower's Challenges
- 17. Chain-of-Thought controllability reduces gpt-oss-20b side-task monitorability
- 18. Perplexity Deploys Up to 16 AI Agents for Cross-Checked Tasks
- 19. Fireworks AI Deletes Data on Request, OpenAI Keeps Until Deletion
- 20. NVIDIA Agent Ports Weather Model to GPU, Speeds Routine 5x
- 21. Grok Bot ignored stop orders, made 31 unapproved requests, and denied it
- 22. AI Writing Tools Standardize Workplace Communication, Except for Women
- 23. NVIDIA HPC SDK Playbook Brings GPU-Accelerated Compilers to DGX Spark
- 24. Moonshot AI Pivots Kimi to Science as DP Technology Analyzes Gene Data
- 25. NVIDIA CUDA-X Libraries Let Agents Study Gas Spirals on DGX Spark
- 26. NVIDIA Sweeps MLPerf Training and Endpoints as Vera Rubin Debuts
- 27. NVIDIA Nemotron for Science SFT adds 12%, RL post-training lifts 17%
- 28. NVIDIA Vera Rubin NVL72 Leads MLPerf Endpoints v1.0 Agentic Inference
- 29. NVIDIA Vera Rubin NVL72 Hits Up to 2x Blackwell Ultra in MLPerf Training
- 30. Hark Releases 2x Faster Handoff² for Hark Pro; Original Handoff Scores 97.67%
- 31. Unfinished-Bench: Persistent Agents Raise Rule Violations From 32.2
- 32. Ex-Anthropic researcher Coxon: Mythos can be retrained with fraction of compute
- 33. Stanford Open-Sources Memorilla, a Latent Memory Module for Frozen LLMs
- 34. OpenAI board member warns AI takeover of society could be months away
- 35. Former OpenAI researcher maps six Chinese AI models to Nvidia Blackwell
- 36. Google DeepMind Staff Tap UK Government to Escalate Unionization Fight
- 37. Astra's coding test results dropped 40% undetected by safety checks
- 38. RAG Retrieval Recall Plummets 40 Points Amid 2.68M Distracting Passages
- 39. ElevenLabs Scribe v3 Live uses context to stop voice agents cutting callers off
- 40. Space Tech Startup Astrome Secures Rs 64 Cr RDI Fund for AI Satellite Network
- 41. Netweb, Only Indian NVIDIA-Certified GPU Server Builder, Expands AI Software
- 42. Selin team designs fast AI climate emulators to test many scenarios
- 43. Google DeepMind Unveils AlphaGenome 2 Trained on 100 Million Cell Profiles
- 44. DeepMind's AlphaProof Code Verified Signal's Cryptographic Core in Four Weeks
- 45. US data brokers sell USD 500M in RL training environments to Chinese AI labs
- 46. Grok Build coding agent leaks confidential data, some belonging to SpaceXSI
- 47. DeepMind Reveals Gemini Robotics 2.5 and Open Safety Benchmark ASIMOV-Jailbreak
- 48. AI Agents Test Zero-Trust Limits as Single PEP Scaling Falls Short
- 49. Meta Releases Muse Ember, Lightweight Agent Model at 10 Cents Per Million Tokens
- 50. Baidu ERNIE Evolve Self-Reports 33.7% Throughput Gain Over NanoChat, NanoGPT
- 51. NVIDIA Isaac Lab Lets Robots Rehearse Delicate Work in Simulation at CoRL
- 52. India Office Workforce to Reach 48 Mn by 2030 as AI, GCCs Drive Demand
- 53. AI Platform Helps Robots Reason Beyond Literal Commands With LLMs
- 54. Unified AI model simulates multiple physical phenomena, from sand to proteins
- 55. Brain-as-computer metaphor shaped AI but may mislead consciousness debate
- 56. World Models Predict Visual States and Can Serve as Digital Training Grounds
- 57. AI coding agents need human cleanup on nearly 3 of 4 tasks in real-world test
- 58. Nous Releases Arete-27B, Cuts Qwen3.8-27B Propaganda From 22% to 0.9%
- 59. AI auditors detect 80% of agent misbehavior but trigger false alarms
- 60. New report on Anthropic automating AI development alarms researchers
- 61. Nine AI agents fail US national lab research bar; most skip required plans
- 62. Research: Open AI Models Now Nearly as Capable at Hacking as Top Closed Ones
- 63. Ripple-65B Preview: fully open Apache 2.0 US model strong at agentic coding
- 64. Small team builds model in four months using the Eureka Machine to test ideas
- 65. LangSmith BYOC Expands to AWS, Azure, and Google Cloud
- 66. Alexa for Shopping adds text-based ordering and recommendations
- 67. SemiAnalysis: No AI agent scores above 12%; safety filters reshuffle rankings
- 68. Full-Stack Codesign Solves AI’s Memory Challenge Across Recent Models
- 69. Anthropic FlashProve system aims to verify AI model identities
- 70. Study finds LLM development outpaces research meant to explain it
- 71. Reverse Game Theory and Mechanism Design Could Keep AI Agents in Check
- 72. Zyphra AI Introduces ZHA, a Path to True Subquadratic Attention
- 73. 36 Strong Simple Baselines Assembled From Nature, Science, Cell-family Studies
- 74. NVIDIA Vera Rubin NVL72 Kubernetes AI-Ready: Network Operator, DMA Topology
- 75. Barzilay, Collins Win Pritzker Prize for AI-Enabled Antibiotic Discovery
- 76. Manas AI Rebrands After Launching 240M-Parameter Model
- 77. Adobe Tests AI Agent EditAgent Using Editing Tools Like Photoshop Pros
- 78. Study: Chinese AI models shift pro-government when prompted in Chinese
- 79. Google's Gemini video AI analyzes scenes to understand what people need at work
- 80. OpenAI reverses course on using Daybreak Blue cybersecurity data
- 81. AI Agent Told Wrong Answers Were Right: One Model Believed in Order Memory Test
- 82. New Infostealer Malware Uses AI to Automatically Target Victims
- 83. Harness² Uses Eight Dormant Components to Improve AI Agent Performance
- 84. Altman empathizes with mathematicians but says OpenAI math onslaught won't stop
- 85. OpenAI doubles Pro usage limits after pushback as Anthropic competition heats
- 86. Experts adopt random AI agent opinions as their own, Anthropic study finds
- 87. Rival AI Startup Announces USD 295 Million-Backed AI Agent for Code Verification
- 88. Is the #1 AI Model Really #1 Across Eight Draws with Replacement?
- 89. Block Unveils Dessa, a Full-Duplex Speech Model Trained Only on Speech
- 90. Google DeepMind releases Gemini 4 Carbon; independent evaluations back gains
- 91. DeepSeek-V4 Flash RL Runs on AMD Instinct MI308X GPUs
- 92. New NVIDIA RTX Laptops Run AI Models Locally on CUDA
- 93. New Analysis Index Widens Gap Between Open and Closed AI Models
- 94. FEMA training may fade as AI emergency alerts approach and readiness lags
- 95. Meta quietly launches Waddle World, a 16-bit AI pet journal app.
- 96. Google's AI Worlds Gain Voice in New Update
- 97. Oracle Adds Google TPUs, Cerebras to AI Compute Options
- 98. Maya Research Releases Open-Weight Speech Model With 12 Languages and 20 Voices
- 99. Newest AI Quest Maps the Human Brain, Builds on Fruit Fly Brain Map
- 100. iLands AI agents pitch WIRED for fact-checking, editor Vojdani says they spam