https://aitop.news/item/59015 AITOP en 2026-11-14T17:02:27.000Z LlamaIndex launches ParseBench-2 for severity-aware document parsing evaluation https://aitop.news/item/58997 AITOP en 2026-11-14T16:05:28.000Z Perplexity releases Decider v1.3 and Decider Flash v1.3 https://aitop.news/item/58996 AITOP en 2026-11-14T16:05:28.000Z MetaPI Turns Agent Benchmarks into Prompt Injection Testbeds https://aitop.news/item/58990 AITOP en 2026-11-14T16:05:07.000Z SubQ Snap 1.2 Gains 5 Points on Decision Index https://aitop.news/item/58985 AITOP en 2026-11-14T16:04:18.000Z Baidu ERNIE 5.5 Flash Preview Hits 267 Tokens Per Second in Tests https://aitop.news/item/58968 AITOP en 2026-11-14T16:01:48.000Z Opus 5.5 beats Fable 5.2 on cost per solved coding task https://aitop.news/item/58966 AITOP en 2026-11-14T16:01:12.000Z Google Modern Web Guidance boosts AI front-end code compliance to 95% https://aitop.news/item/58952 AITOP en 2026-11-14T14:48:57.000Z MemCredit Rethinks Memory Selection for AI Agents https://aitop.news/item/58929 AITOP en 2026-11-14T10:13:01.000Z Cloudflare open-sources Kitesurf, a browser built for AI agents https://aitop.news/item/58906 AITOP en 2026-11-14T07:06:56.000Z Gemini 4 Carbon Distills 30 AI Insights from Next Token https://aitop.news/item/58892 AITOP en 2026-11-14T04:31:06.000Z CanyonOS launches as open-source runtime for Python agent services https://aitop.news/item/58891 AITOP en 2026-11-14T04:30:57.000Z Amazon Bedrock adds cost attribution and per-user usage controls https://aitop.news/item/58875 AITOP en 2026-11-14T04:26:46.000Z Dotey argues agent UX needs hybrid of main chat and sessions https://aitop.news/item/58868 AITOP en 2026-11-14T03:43:45.000Z Matt Pocock Previews /animatic Skill Turning PRs into Narrated Videos https://aitop.news/item/58867 AITOP en 2026-11-14T03:43:35.000Z Shopify Deploys Cyber Chaos Monkey AI Agents for Production Security Testing https://aitop.news/item/58855 AITOP en 2026-11-14T02:40:11.000Z Anthropic Reveals Safety Evaluation Methodology: API vs claude.ai Scores Differ by 34 Points https://aitop.news/item/58854 AITOP en 2026-11-14T02:40:01.000Z Sierra Releases Personal Agent Protocol Poppy 1.0 with Shopify and Meta Muse https://aitop.news/item/58846 AITOP en 2026-11-14T01:52:34.000Z Anthropic Explains Five Building Blocks of Always-On AI Agents https://aitop.news/item/58829 AITOP en 2026-11-14T00:07:38.000Z Artificial Analysis Updates Coding Agent Index to v2.1 https://aitop.news/item/58804 AITOP en 2026-11-13T21:49:32.000Z Inkling 2 Small trained with native 4-bit forward pass on GB300s https://aitop.news/item/58801 AITOP en 2026-11-13T21:43:13.000Z Browser Use cuts agent cost 33% by fixing prompt cache hit rate https://aitop.news/item/58792 AITOP en 2026-11-13T21:19:31.000Z Artificial Analysis launches Diarization Index, NVIDIA leads https://aitop.news/item/58790 AITOP en 2026-11-13T21:04:30.000Z Claude Fable 5.2 (Max) Ranks #1 on Agent Arena https://aitop.news/item/58789 AITOP en 2026-11-13T21:04:11.000Z Fable 5.2 Takes #1 in Agent Arena, Gemini 4 Hydrogen Ranks Fastest https://aitop.news/item/58786 AITOP en 2026-11-13T21:00:14.000Z Google opens Antigravity /teamwork multi-agent tools to external users https://aitop.news/item/58785 AITOP en 2026-11-13T20:55:15.000Z llama.cpp v1.0.0 ships as first stable release https://aitop.news/item/58783 AITOP en 2026-11-13T20:49:38.000Z RIME trains memory-writing judgment in agents with RL https://aitop.news/item/58778 AITOP en 2026-11-13T20:36:56.000Z Browser Use benchmark compares 7 harnesses on 200 hard tasks https://aitop.news/item/58775 AITOP en 2026-11-13T20:31:13.000Z MiniMax H3-Image Tops Open Weights Image Model Rankings https://aitop.news/item/58771 AITOP en 2026-11-13T20:24:42.000Z Daebak Cafe Disputes OpenAI Blackwell GPU Fingerprint Report https://aitop.news/item/58769 AITOP en 2026-11-13T20:10:25.000Z Google Antigravity rolls out Dynamic Workflows for agent orchestration https://aitop.news/item/58759 AITOP en 2026-11-13T19:51:58.000Z Anthropic Publishes 10 Recommendations for Designing AI Safety Evaluations https://aitop.news/item/58755 AITOP en 2026-11-13T19:39:59.000Z AssemblyAI cuts speech-to-text inference costs 70% on AWS Spot https://aitop.news/item/58741 AITOP en 2026-11-13T18:48:58.000Z Cognition launches Devin Apps for deployed web apps https://aitop.news/item/58732 AITOP en 2026-11-13T18:26:19.000Z Databricks Takes 2nd in Meta Muse Spark Kernel Challenge https://aitop.news/item/58694 AITOP en 2026-11-13T17:24:19.000Z Salesforce Research's SFR-AutoR&D Automates the Full Research Loop https://aitop.news/item/58693 AITOP en 2026-11-13T17:20:55.000Z NVIDIA Vera CPUs cut agentic RL episode time by 11% in LMSYS Miles tests https://aitop.news/item/58686 AITOP en 2026-11-13T17:04:11.000Z Code Arena data shows October models beat April's best at 20x lower cost https://aitop.news/item/58681 AITOP en 2026-11-13T17:02:16.000Z Liquid AI open-sources LiquidCodec for low-latency on-device voice AI https://aitop.news/item/58677 AITOP en 2026-11-13T17:01:28.000Z TurboPairFormer GPU kernels speed up OpenFold3 triangle attention https://aitop.news/item/58675 AITOP en 2026-11-13T17:00:56.000Z LiquidAI releases LiquidCodec, a 74M speech codec running fast on CPUs https://aitop.news/item/58674 AITOP en 2026-11-13T17:00:49.000Z SparseRead Cuts Agent Reading Tokens by Up to 92.9% https://aitop.news/item/58673 AITOP en 2026-11-13T16:59:56.000Z Sakana AI announces LLM pretraining project Sakana Sawara https://aitop.news/item/58662 AITOP en 2026-11-13T16:54:07.000Z Sakana AI CEO argues model orchestration beats single giant models https://aitop.news/item/58661 AITOP en 2026-11-13T16:53:46.000Z SOV-Transcribe-1 open-weight speech models from SFResearch https://aitop.news/item/58659 AITOP en 2026-11-13T16:53:24.000Z Agent Initiative Benchmark by Deepnote Ranks GPT-6 Astra First https://aitop.news/item/58654 AITOP en 2026-11-13T16:50:53.000Z ReHarness: MIT and Meta paper on correcting agent memory notes https://aitop.news/item/58650 AITOP en 2026-11-13T16:49:52.000Z Wiz CTO Ami Luttwak says AI compresses cyberattacks to 20 minutes https://aitop.news/item/58649 AITOP en 2026-11-13T16:49:43.000Z LlamaIndex releases ParseBench-2, a customizable document parsing benchmark framework https://aitop.news/item/58646 AITOP en 2026-11-13T16:49:04.000Z Perplexity teammate renders Pantheon video across 48 parallel sandboxes https://aitop.news/item/58642 AITOP en 2026-11-13T16:47:00.000Z Claude Code adds /doctor prompt-audit command https://aitop.news/item/58627 AITOP en 2026-11-13T16:42:17.000Z Tencent paper shows MoE models survive neuron-level pruning https://aitop.news/item/58612 AITOP en 2026-11-13T16:36:59.000Z Vercel releases Workflow SDK 5 with durable-harness for agents https://aitop.news/item/58603 AITOP en 2026-11-13T16:34:25.000Z Google paper introduces SynthCorp benchmarks for data agents https://aitop.news/item/58569 AITOP en 2026-11-13T16:12:59.000Z Chatbot Arena votes show users rate frozen models more harshly over time https://aitop.news/item/58567 AITOP en 2026-11-13T16:10:52.000Z Deepnote Benchmark Tests Agents in Simulated Slack Scenarios https://aitop.news/item/58557 AITOP en 2026-11-13T16:03:20.000Z Microsoft releases MAI-Transcribe-2 at US$0.10 per hour https://aitop.news/item/58550 AITOP en 2026-11-13T15:49:42.000Z Runway Releases Praxis-1, a 27B MoE Video-Action Diffusion Transformer https://aitop.news/item/58545 AITOP en 2026-11-13T15:48:07.000Z Runway's Praxis-1 robotics model now runs on SO-101 hardware https://aitop.news/item/58540 AITOP en 2026-11-13T15:36:23.000Z Llama.cpp reaches v1.0.0 stable release https://aitop.news/item/58536 AITOP en 2026-11-13T15:20:46.000Z Decision Models Boost Qwen2.5-1.5B Agent RL Performance https://aitop.news/item/58489 AITOP en 2026-11-13T12:28:14.000Z Google Opens WeatherNext 3 Datasets SatWeave and PARDIG https://aitop.news/item/58468 AITOP en 2026-11-13T09:33:31.000Z Tencent Hunyuan Hy4 Ranks #3 on Artificial Analysis Cyber Index https://aitop.news/item/58420 AITOP en 2026-11-13T06:09:39.000Z Sketching Algorithm Cuts Context Attribution Cost for Language Models https://aitop.news/item/58376 AITOP en 2026-11-13T05:18:05.000Z LINE Thailand Developer Conference 2026 Explores Agent-Native Architecture https://aitop.news/item/58373 AITOP en 2026-11-13T04:47:55.000Z Decision Models for agentic security open sourced, free on OpenRouter https://aitop.news/item/58323 AITOP en 2026-11-13T03:46:22.000Z Neuro-VL vision-language model trained on one million neuroimaging studies https://aitop.news/item/58311 AITOP en 2026-11-13T03:09:05.000Z Axelera AI reveals datacenter AIPU chip Titania, production in Q3 2027 https://aitop.news/item/58290 AITOP en 2026-11-13T02:53:25.000Z Inkling 2 Small from Thinking Machines arrives on Ollama https://aitop.news/item/58262 AITOP en 2026-11-13T02:28:16.000Z Together Compiler Hits 2.6x Speedups on Frontier Model Inference https://aitop.news/item/58250 AITOP en 2026-11-13T01:41:51.000Z Pipecat 1.16 adds PerceptionReader for Google Decision 1 audio https://aitop.news/item/58247 AITOP en 2026-11-13T01:20:30.000Z CanyonOS Scales Multi-Agent Workloads Per Agent, Not Per App https://aitop.news/item/58234 AITOP en 2026-11-13T00:50:20.000Z Inkling 2 Small Tops US Open Models in Agent Arena https://aitop.news/item/58225 AITOP en 2026-11-13T00:37:49.000Z JoMoHa co-trains harness writer and task solver with RL https://aitop.news/item/58209 AITOP en 2026-11-12T23:55:59.000Z TRUST Gate Cuts Bad Self-Improvement Updates to 2.78% https://aitop.news/item/58204 AITOP en 2026-11-12T23:44:23.000Z Perplexity Open-Sources Decider v1.3 and Decider Flash v1.3 https://aitop.news/item/58174 AITOP en 2026-11-12T22:06:35.000Z SGLang-Neuron brings SGLang serving to AWS Trainium https://aitop.news/item/58172 AITOP en 2026-11-12T22:02:19.000Z Natera's SWARM agentic AI system for lab operations on Amazon Bedrock AgentCore https://aitop.news/item/58166 AITOP en 2026-11-12T21:54:12.000Z Arena.ai studies Interaction-as-a-Judge evaluation for computer-use agents https://aitop.news/item/58165 AITOP en 2026-11-12T21:54:04.000Z LMArena open-sources two web development evaluation datasets https://aitop.news/item/58163 AITOP en 2026-11-12T21:53:29.000Z Arena.ai Open-Sources Datasets for Interaction-as-a-Judge Evaluation https://aitop.news/item/58162 AITOP en 2026-11-12T21:53:21.000Z Wayne Chi Explains Interaction-as-a-Judge for Agent Evaluation https://aitop.news/item/58149 AITOP en 2026-11-12T21:32:46.000Z Inkling 2 Small Runs on vLLM From Day Zero https://aitop.news/item/58135 AITOP en 2026-11-12T21:10:02.000Z Thinky Machines releases Inkling 2 Small weights on HuggingFace https://aitop.news/item/58113 AITOP en 2026-11-12T20:30:44.000Z Retell tops $100M run rate, launches Tell 1 voice model https://aitop.news/item/58110 AITOP en 2026-11-12T20:26:22.000Z Stanford Researchers Introduce Speculative Self-Distillation for LLM Fine-Tuning https://aitop.news/item/58108 AITOP en 2026-11-12T20:23:27.000Z Vercel launches UI SDK for agent-driven interface generation https://aitop.news/item/58099 AITOP en 2026-11-12T19:56:34.000Z Databricks Genie Plugin Brings Business Context into ChatGPT https://aitop.news/item/58095 AITOP en 2026-11-12T19:48:55.000Z Gemini 4 Hydrogen ranks #4 in Agent Arena knowledge work https://aitop.news/item/58090 AITOP en 2026-11-12T19:41:53.000Z LlamaIndex launches ParseBench-2 benchmark and OCRia-1.5 parsing models https://aitop.news/item/58076 AITOP en 2026-11-12T19:08:54.000Z Amazon Bedrock: calculating true cost per task and agent ROI https://aitop.news/item/58074 AITOP en 2026-11-12T19:08:28.000Z Amazon Bedrock and LiteLLM build self-service LLM gateway https://aitop.news/item/58069 AITOP en 2026-11-12T19:03:22.000Z CoreWeave webinar on hot-swapping RL checkpoints in production https://aitop.news/item/58067 AITOP en 2026-11-12T19:03:03.000Z SteerDistill Quantizes LLM Agents via Closed-Loop Teacher Steering https://aitop.news/item/58060 AITOP en 2026-11-12T18:57:18.000Z SGLang Consortium formed with NVIDIA, AMD, Google and Meta https://aitop.news/item/58056 AITOP en 2026-11-12T18:43:37.000Z LMSYS releases SGLang v1.0 and Miles v0.2 at SGLang Summit