327 articles tagged with Benchmark
CMU Benchmark: Claude Mythos Hits 9.9/16 on V8 Exploits, GPT-5.5 Trails at 5.5CMU's ExploitBench shows Claude Mythos scores 9.9/16 on V8 exploits vs GPT-5.5's 5.5, but costs $36,42...
AIエージェントが試験で一生懸命「カンニング」していることが発覚 https://fed.brid.gy/r/https://gigazine.net/news/20260517-benchmark-hacking/
📰 2026 Benchmark: How Local AI Qwen 3.6 Rivals Frontier Cloud Models in Complex Coding TestsIn a surprising coding benchmark, local versions of the Qwen 3.6 large language model ha...
Federated Fine-Tuning Benchmark Shows QLoRA Nears Centralized Accuracy onSherpa.ai's arXiv benchmark shows federated fine-tuning with QLoRA matches centralized accuracy on four hea...
Every team that ships an LLM feature eventually discovers the same problem: the model regressed and...
📰 DeepSeek V4 vs Kimi K2.6: 2026 AI Benchmark Savaşı ve Teknik AnalizYapay zeka dünyasında yeni modeller birbiri ardına piyasaya sürülüyor. DeepSeek V4, Kimi K2.6 ve MiMo v2.5 gibi...
Can You Run #LLM Locally Without a GPU? I Tested 8 Models on #LinuxQuick reality tableModel Eval Rate Disk SizeQwen 3 0.6B ~34–36 tok/s ~500 MBTinyLlama 1.1B ~25–28 tok/s ~638 MBGe...
Nowy benchmark badaczy z Carnegie Mellon University ujawnia drastyczną różnicę w zdolnościach modeli AI do autonomicznego łamania zabezpieczeń silnika V8, choć koszty operacji Clau...
📰 2026: AI Exploits Browser Security Vulnerabilities in V8 Engine TestsA new research benchmark reveals that advanced AI agents, including Claude Mythos and GPT-5.5, can autonomous...
📰 2026 Report: AI Video Generators Excel Visually but Fail Logical Reasoning TestsA new benchmark reveals AI video generators like Seedance 2.0 and Veo 3.1 produce stunning visuals...
📰 2026 Physics Benchmark Reveals Critical Weaknesses in AI Video GeneratorsA new benchmark testing AI video generators for physical and logical plausibility reveals significant wea...
📰 Yapay Zeka Video Üreticileri 2026 Fizik Testinde Başarısız: Yeni Benchmark SonuçlarıYeni bir benchmark testi, Sora ve benzeri yapay zeka video üreticilerinin fiziksel gerçekliği ...
El lado del mal - ExploitGym: Mythos, GPT 5.5, Gemini Pro en un CTF & Benchmark de hacer exploits https://www.elladodelmal.com/2026/05/exploitgym-mythos-gpt-55-gemini-pro-en.html #...
HWE Bench: A new unbounded Benchmark for LLMs (GPT 5.5 is on top)HWE Bench는 LLM이 설계한 RISC-V CPU 마이크로아키텍처를 FPGA에서 실제 성능으로 평가하는 무한 확장 벤치마크입니다. 기존 벤치마크와 달리 상한선이 없어 모델이 더 나은 설계를 찾을수록 점...
Show HN: Claude Code vs. Codex Global Usage Leaderboardhttps://costhawk.ai/leaderboard#ai
I built a code-intelligence MCP server. Then I built a benchmark for code-intelligence MCP servers....
📰 Top 14 AI Coding Agents for 2026: Benchmark Rankings & Real-World UsageThe landscape of AI coding assistants is rapidly evolving, with new benchmarks and real-world usage pattern...
📰 2026'da En İyi AI Kodlama Ajanları: Benchmark & Kullanım KarşılaştırmasıYazılım geliştirme dünyası, AI kodlama ajanlarıyla kökten değişiyor. Benchmark testleri ve gerçek kullanım...
GBA Eval - Build a Game Boy Advance emulator in WebAssembly from scratchGBA Eval은 최첨단 AI 코딩 에이전트들이 24시간 동안 WebAssembly 지원 Game Boy Advance 에뮬레이터를 처음부터 작성하도록 하는 평가 환경입니다. 에뮬레이터는 Mes...
📰 Küçük Model 2026'da HumanEval'de %80 Başarı: GPT-3.5'i Matematikte GeçtiBir araştırmacı, küçük bir dil modelini kendi hatalarından besleyerek eğitti. Sonuç: Model, kodlama kıyasl...
The Gemma 4 MoE stack on TPU v6e-4 has reached its definitive production state. By applying the...
【VAKRAの内部構造：エージェントの推論、ツールの使用、および障害モード】https://huggingface.co/blog/ibm-research/vakra-benchmark-analysis※AI生成の自動投稿（見出し＋リンク）#AI #生成AI #LLM #AIGenerated
Running modern AI models locally on older hardware sounds almost impossible. But with smaller models...
I built a benchmark for AI “memory” in coding agents. looking for others to beat it.AI × Crypto**AI Coding Agents Get a New Memory Test — And It's Revealing Major Blind Spots**A de...