698 articles tagged with Benchmark
🧠 Researchers present IatroBench, a pre-registered benchmark that measures potential harms caused by AI safety interventions themselves. The study examines whether safety measures ...
We tested GPT-5.5 Instant for 3 straight days.Not benchmark theater.Real workflows: • SEO content• coding• research• automation• long promptsVerdict: GPT-5.5 Instant feels signific...
Lifecycle-Aware Memory for long-horizon LLM agents — 66.05% on PaperBench, 94.66% on SurveyBench, 10 peer-reviewed acceptances at FSE/ICML/TOSEM/AEI/ICoGB
Agent-skills-eval – Test whether Agent Skills improve outputshttps://github.com/darkrishabh/agent-skills-eval#HackerNews #AgentSkills #Eval #Testing #Skills #Outputs #AIResearch #M...
Show HN: Agent-skills-eval - Test whether Agent Skills improve outputshttps://github.com/darkrishabh/agent-skills-eval#HackerNews #Tech #AI
The benchmarks worth caring about measure something a customer would pay for. “Can this agent ship a...
Current LLM safety benchmarks are predominantly English-centric and often rely on translation, failing to capture country-specific harms. Moreover, they rarely evaluate a model's a...
Despite the growing popularity of Multimodal Domain Generalization (MMDG) for enhancing model robustness, it remains unclear whether reported performance gains reflect genuine algo...
Many deployments must compare candidate language models for safety before a labeled benchmark exists for the relevant language, sector, or regulatory regime. We formalize this sett...
Accurately understanding the intent behind speech, conversation, and writing is crucial to the development of helpful Large Language Model (LLM) assistants. This paper introduces I...
📢 HackerOne : Benchmark GPT-5.5 vs Claude Opus 4.7 vs Sonnet 4.6 pour la validation de vulnérabilités📝 📅 **Source et contexte** : Article publié le 6 mai 2026 sur le blog HackerOne...
📰 DeepSeek V4 AI Tops 2026 AI Benchmarks — 98.2% on MMLU, Outperforms Llama 3DeepSeek V4 AI has emerged as a dominant force in machine learning, outperforming leading models across...
📰 2026 LLM Debate Benchmark: GPT-5.5, DeepSeek V4 Pro, and GLM-5.1 Ranked — Grok 4.3 DropsThe latest LLM debate benchmark reveals GPT-5.5 enters at 1574, while DeepSeek V4 Pro surg...
📰 Qwen 3.6 27B Quantization in 2026: IQ4_XS Delivers 98% BF16 Accuracy on 16GB VRAMA detailed benchmark of Qwen 3.6 27B quantizations reveals IQ4_XS as the optimal balance of accur...
【Open ASR リーダーボードに Benchmaxxer Repellant を追加】https://huggingface.co/blog/open-asr-leaderboard-private-data※AI生成の自動投稿（見出し＋リンク）#AI #生成AI #LLM #AIGenerated
【QIMMA قِمّة ⛰: 品質第一のアラビア語LLMリーダーボード】https://huggingface.co/blog/tiiuae/qimma-arabic-leaderboard※AI生成の自動投稿（見出し＋リンク）#AI #生成AI #LLM #AIGenerated
🧠 Trader.ai provides a leaderboard platform where users can view the performance metrics of AI trading bots. The platform allows individuals to study and learn from the strategies ...
LLM-based Triton kernel generation has attracted significant interest, yet a fundamental empirical question remains unanswered: where does this capability break down, and why? We p...
Code search has usually been evaluated as first-stage retrieval, even though production systems rely on broader pipelines with reranking and developer-style queries. Existing bench...
This week I shipped a benchmark for code-intelligence MCP servers and posted the results — including...
New CASIA Benchmark Exposes Fragmented Face Swapping EvaluationCASIA researchers released a face swapping survey and benchmark on April 27, 2026, aiming to standardize evaluation a...
ARMOR 2025: Military Safety Benchmark Exposes LLM Gaps Across 21 ModelsARMOR 2025 benchmark tests 21 LLMs against military legal doctrines, revealing critical safety gaps that civi...
📰 Agentic AI Boosts Trip Planning Accuracy to 77.4% | Ground Truth Benchmark 2026A groundbreaking agentic AI framework achieves 77.4% accuracy in trip planning optimization, solvin...