Havuz

⚡ DROS™ VajraClaw (Hacker Edition): Free Multi-Agent Runtime Security & Docker Governance Gateway with W3C DID & <1μs Enforcement

[NeurIPS 2026 Meta-Agents] PostTrainBench⁰ — Can LLM agents automate LLM post-training without gradients?

A JAX transport-economy learning environment for route planning, cargo flow, financing, and replayable agent benchmarks.

Measures which GGUF quant of a model to run on your GPU: real speed and quality loss (KL divergence), not estimates.

Cognitive-structured Multimodal Agent (CMA-Harness): a memory-centric agent for long-horizon multimodal understanding, generation, and editi

Measures how much code LLM agents waste. 5 tasks, sandboxed, weekly snapshots, contestable references.

StealthBench is a benchmark for evaluating whether autonomous agents operate with the operational security (OPSEC) discipline that separates

Repository for the implementation of MacArena

⚽🤖 11 frontier LLMs predicted the entire 2026 World Cup — frozen before kickoff. Live leaderboard: Brier score, bracket points & Polymarket

Eval framework that ranks agentic coding frameworks (Superpowers, Compound Engineering, Agent Skills, GSD) by having each build the same spe

Benchmark llama.cpp on your own GPU/CPU boxes and compare against everyone else's — sweep runner, an sha256-keyed index of ~4M GGUF files, a

An objective reconstruction evaluation toolkit for audio tokenizers, neural codecs, audio VAEs, and vocoders

Lightweight agent that pulls and runs LLM benchmarks from Protorikis Bench

Red team vs blue team benchmark for scientific paper verification. Plant defects into open access biomedical papers, build gates to catch th

同一份提示词，两座汴京：Claude Opus 5.5 与 GPT 6 Sol 各自从零写出的《清明上河图》实时 3D 网页 · Same prompt, two cities — a real-time 3D Bianjing benchmark for LLMs

Baba Is Harbor - AI agent benchmark on Baba is You

English, Japanese, and Chinese LLM safety mini-benchmark.

A Open-Source Search Feature.

OpenSkillEval: Automatically Auditing the Open Skill Ecosystem for LLM Agents

SciPIvot: Benchmarking Principal-Investigator Judgment at Pivotal Moments of Scientific AutoResearch Agents

OptiScaler ILION — Community fork with bilingual ES/EN interface, FSR 4 version switching, live performance metrics and benchmarking tools.

Execution Before Memory: diagnosing memory failures in video world models.

AnyFlow - flow-matching video generation: training, inference and VBench evaluation for t2v and i2v models.

Ứng dụng miễn phí kiểm tra sức khoẻ ổ đĩa & hệ thống cho macOS — SMART, nhiệt độ, benchmark tốc độ, phát hiện USB/SSD giả dung lượng. Cho cộ

A temporal visual reasoning benchmark over egocentric robot manipulation video — 39 scenarios, 7 dimensions, 34,713 questions

A benchmark of AI QA agents finding real bugs in 4 open-source web apps, with 1,396 labeled bugs.

Streaming privacy for LLM apps: pii-leak-benchmark (does your gateway leak PII in streams?), chunk-invariance (one-line test for split-bound

Did your change make your LLM server faster? Measure it: faster, slower, or too close to call. Linux and Apple Silicon.

Benchmark for repository-level code understanding: 500 open-ended questions from real GitHub Discussions across 50 repositories, graded by a

A benchmarking suite for Agent Skills (SKILL.md files) that measures whether a skill actually improves LLM output quality, and by how much.

Reproducible benchmark that compares VictoriaLogs, Loki, Elasticsearch, OpenSearch, SigNoz and ClickStack

MiniMax H3: 4.9x faster local generation and seam-free 4K via synchronized tiles (SyncTile). Benchmarks included.

Open-source benchmark for evaluating LLMs on domain-specific geological reasoning, AI for geology, and source-grounded geological knowledge.

Layang is a gRPC-first, workspace-based API workbench for testing, mocking, benchmarking, documenting, and automating API workflows across g

Compressed, action-first output style for Claude Code — a benchmarked merge of caveman and i-have-adhd, scored on quality as well as tokens.

A measurement-grounded benchmark for evaluating physical fidelity in physics engines and video world models.

Several coding agents, one repo, at the same time: a reproducible lab to measure how they coordinate, and Médula, a kernel that decides who

What can Jev actually do? Reproducible experiments and research reports exploring its capabilities and limits.

Measures how much code AI coding agents add, from patches benchmarks already publish

Network benchmarking suite for DNS, HTTP, and SSL — with DNSSEC validation, DoH/DoT support, high‑concurrency testing, and enterprise‑grade

An arena where LLMs write robot controllers and their code competes.

Hardware-in-the-loop benchmark for agentic embedded AI deployment. Agents iteratively edit model or firmware artifacts, the framework compil

CatchBench: a benchmark for auditing agent failures across the full lifecycle.

Project page for AV-Phys Bench: Do Joint Audio-Video Generation Models Understand Physics?

Don't trust an Agent. Test it. Open-source lab for agent credit — exams → evidence → explainable, recomputable scores. Built for the day age

curl performance tests

Open browser arena for embodied / VLA stacking benchmarks

Catches multilingual voice agents placing the wrong order. τ-bench style benchmark + CI gate for voice ordering agents in English and Hingli

J-Space Cognition Suite, AI agent görevlerinde akıl yürütme, çok aşamalı iş takibi ve doğrulamayı yönlendirmek için Skill biçiminde paketlenmiş bir inference-time kontrol yaklaşımıdır.

Neden yükseliyor · 42 günde 1026 yıldız alması ve GLM-5.3-Flash × J-Space capability sunumuyla ilişkilendirilmesi dikkat çekiyor. Son 28 günde 18 commit görünse de bunların tek yazardan gelmesi, bağımsız katkı veya benimseme kanıtını sınırlıyor; mevcut bilgiler yükselişin nedenini ya da iddia edilen kabiliyetleri doğrulamıyor.

A unified framework for training, fine-tuning, and evaluating World Action Models