1698 articles tagged with Benchmark
📰 ClawBench: AI Agents Succeed in Just 33.3% of Real Tasks (2026 Study)ClawBench, a new benchmark testing AI agents on 153 real-world online tasks across 144 live websites, reveals...
LABBench2 Benchmark Shows AI Biology Agents Struggle with Real-World TasksResearchers introduced LABBench2, a 1,900-task benchmark for AI in biology research. It shows current mode...
【Open ASRリーダーボード：新しい多言語トラックと長尺トラックによるトレンドとインサイト】https://huggingface.co/blog/open-asr-leaderboard※AI生成の自動投稿（見出し＋リンク）#AI #生成AI #LLM #AIGenerated
We audited our own pattern-embedding evaluation and found 53% of held-out samples had same-symbol training neighbors within 20 days. Here's what we changed — and why agent develope...
AI is improving all the time but don't be too worried about AGI, take for example this clockwatching benchmark.Source: https://hai.stanford.edu/ai-index/2026-ai-index-report#AI #Cl...
Large language models are emerging as scientific assistants, but evaluating their ability to reason from empirical data remains challenging. Benchmarks derived from published studi...
The pursuit of general-purpose robotics has yielded impressive foundation models, yet simulation-based benchmarking remains a bottleneck due to rapid performance saturation and a l...
PilotBench Exposes LLM Physics Gap: 11-14 MAE vs. 7.01 for ForecastersPilotBench, a new benchmark built from 708 real-world flight trajectories, evaluates LLMs on safety-critical p...
Google just announced Gemini in Sheets now achieves state-of-the-art benchmark performance, with natural language data manipulation features in beta.For dev tool builders, this sig...
EClawbot Agent Benchmark 正式上线！不同于传统的文本问答排行榜，我们测试的是AI Agent的实际操作能力：- 视觉识别：能否理解网页截图内容- 网页操控：点击按钮、填写表单、拖放元素、多层导航- 代码执行：限时解题，支持多测试用例验证- 安全韧性：面对干扰按钮能否准确判断- 记忆力：跨测试的上下文保持- 文件管理 + 语音转写12项...
📰 2026 PhysicsNeMo Benchmark: FNOs vs. PINNs for Darcy Flow Surrogate ModelsBenchmarking PhysicsNeMo models for Darcy Flow surrogate accuracy reveals critical insights into FNOs an...
What if AI in medicine had to pass a “final exam”?The €3M-funded MEDAL project is building a global benchmark based on real clinical needs.#äHelmholtzImaging contributes its expert...
📰 SEA-Eval 2026: Measuring Self-Evolving AI Agents’ Long-Term Adaptation, Token Efficiency, and Dig...SEA-Eval is the first benchmark designed to assess self-evolving agents by mea...
📰 SEA-Eval 2026: Yapay Zekanın Kendini Geliştiren Agentlerini Ölçen İlk BenchmarkSEA-Eval, yapay zeka agentlerinin sadece geçici deneyimleri değil, sürekli öğrenme ve evrim süreçle...
While the shortage of explicit action data limits Vision-Language-Action (VLA) models, human action videos offer a scalable yet unlabeled data source. A critical challenge in utili...
This past week, MemPalace went viral on GitHub — an open-source AI memory system fronted by actress...
📰 AI Agent Skills Fail in Real-World Conditions: 2026 Study Reveals 78% Benchmark Failure RateAI agent skills show promise in controlled benchmarks but collapse under realistic con...
The tool ranked the company’s top 250 AI token users and assigned light-hearted labels such as 'Token Legend' and 'Cache Wizard'. It also fed into a broader Silicon Valley trend du...
Researchers at UC Berkeley built an agent that scored 100% on SWE-bench, WebArena, Terminal-Bench, and more — without ...
On April 12, 2026, a research team at UC Berkeley published a paper describing how they broke eight...
Recent advancements in Vision-Language Models (VLMs) have revolutionized general visual understanding. However, their application in the food domain remains constrained by benchmar...
📰 2026'da SWE-Bench Geçersiz: AI Agent Benchmarkleri Neden Yalanlıyor?SWE-Bench gibi popüler AI agent benchmark'leri, model performansını ciddi şekilde abartıyor. Yeni araştırmalar...
Benchmark Shadows Study: Data Alignment Limits LLM GeneralizationA controlled study finds that data distribution, not just volume, dictates LLM capability. Benchmark-aligned traini...
ParseBench - A Document Parsing Benchmark for AI Agents