Library
85,762 resources from 11 public sources. 54 checked by hand; the rest are unreviewed.
Results
-
sacp
Datasets & benchmarksState-Aware Collaboration Protocol: auditable work receipts for AI agents.
git clone https://github.com/aDragon0707/sacp.git -
HallusionBench
Datasets & benchmarks[CVPR'24] HallusionBench: You See What You Think? Or You Think What You See? An Image-Context Reasoning Benchmark Challenging for GPT-4V(ision), LLaVA-1.5, and Other Multi-modality Models
git clone https://github.com/tianyi-lab/HallusionBench.git -
openbench
Datasets & benchmarksSame model, different wrapper: a from-scratch benchmark comparing coding-agent harnesses (codex, pi, opencode, cursor, devin) and open models on correctness, speed, and token cost
git clone https://github.com/minghinmatthewlam/openbench.git -
UHGEval
Datasets & benchmarks[ACL 2024] User-friendly evaluation framework: Eval Suite & Benchmarks: UHGEval, HaluEval, HalluQA, etc.
git clone https://github.com/IAAR-Shanghai/UHGEval.git -
TerminalWorld
Datasets & benchmarksTerminalWorld: Benchmarking Agents on Real-World Terminal Tasks (NeurIPS 2026 Evaluations and Datasets Track)
git clone https://github.com/terminalworld/TerminalWorld.git -
Long-Horizon-Terminal-Bench
Datasets & benchmarksHugging Face dataset by IntelligenceLab
IntelligenceLab/Long-Horizon-Terminal-Bench -
mac-ml-speed-test
Datasets & benchmarksA few quick scripts focused on testing TensorFlow/PyTorch/Llama 2 on macOS.
git clone https://github.com/mrdbourke/mac-ml-speed-test.git -
Omni-SafetyBench
Datasets & benchmarks[ACM MM 2026 Oral] Code for paper "Omni-SafetyBench: A Benchmark for Safety Evaluation of Audio-Visual Large Language Models".
git clone https://github.com/THU-BPM/Omni-SafetyBench.git -
meta-agents-research-environments
Datasets & benchmarksMeta Agents Research Environments is a comprehensive platform designed to evaluate AI agents in dynamic, realistic scenarios. Unlike static benchmarks, this platform introduces evolving environments where agents must adapt their strategies as new information becomes available, mirroring real-world challenges.
git clone https://github.com/facebookresearch/meta-agents-research-environments.git -
AIRTBench-Code
Datasets & benchmarksCode Repository for: AIRTBench: Measuring Autonomous AI Red Teaming Capabilities in Language Models
git clone https://github.com/dreadnode/AIRTBench-Code.git -
arena-ai-leaderboards
Datasets & benchmarks📊 Daily auto-updated snapshots of all Arena AI (LMSYS Chatbot Arena) leaderboards — LLM, Vision, Code, Video, Image & more. Structured JSON with historical tracking.
git clone https://github.com/oolong-tea-2026/arena-ai-leaderboards.git -
TwinFix
Datasets & benchmarksTwinFix — AI agent that diagnoses, reproduces and fixes bugs in a disposable twin environment. 在一次性孪生环境中自动诊断、复现并修复故障的 AI Agent。
git clone https://github.com/spacesky-cell/TwinFix.git -
VisualNews-Repository
Datasets & benchmarks[EMNLP'21] Visual News: Benchmark and Challenges in News Image Captioning
git clone https://github.com/FuxiaoLiu/VisualNews-Repository.git -
CodeX-2M-Thinking
Datasets & benchmarksHugging Face dataset by Modotte
Modotte/CodeX-2M-Thinking -
ReForm-Eval
Datasets & benchmarksAn benchmark for evaluating the capabilities of large vision-language models (LVLMs)
git clone https://github.com/FudanDISC/ReForm-Eval.git -
NIMStats
Datasets & benchmarks📊 Automated hourly benchmarks for 20+ NVIDIA NIM models — interactive dashboard, zero infra, self-hostable. Open-source & community-driven.
git clone https://github.com/MauroDruwel/NIMStats.git -
VibeSearchBench
Datasets & benchmarks🔍 The hardest search benchmark in the wild — vague, multi-turn, proactive. 200 long-horizon tasks with persona-driven progressive disclosure, scored by verifiable schema-free knowledge-graph evaluation. No vibes, just triplet F1.
git clone https://github.com/VibeBench/VibeSearchBench.git -
qwen38-strix-halo-harness
Datasets & benchmarksAI coding agent harness for any hardware, local or cloud (OpenAI-compatible engines). Tuned for AMD Strix Halo: Qwen3.8 125B on the iGPU, NPU side-jobs, crash recovery, logged benchmarks.
git clone https://github.com/aic0d3r/qwen38-strix-halo-harness.git -
agent-duel
Datasets & benchmarksBlinded battles between coding agents: two agents, one prompt, you pick the winner. Desktop app, runs locally.
git clone https://github.com/bottomless/agent-duel.git -
multi-agents
Datasets & benchmarksMulti-agent teams for real work: run teams of coding agents (Claude Code, Codex, any model) in isolated git worktrees, and measure whether the team beats a solo agent on cost and results.
git clone https://github.com/BichengWang/multi-agents.git -
primitive-bench
Datasets & benchmarksThe marketplace for verifiable AI outcomes: state a task, get a fixed price upfront, and Primitive Bench routes across tools to complete it, refunded if the outcome isn't delivered. This repo is the open verification layer.
git clone https://github.com/primitive-bench/primitive-bench.git -
xorcise
Datasets & benchmarksRun cyber-AI agents against real missions in isolated environments. Record everything as OpenTelemetry evidence. Grade the evidence, not the claim.
git clone https://github.com/xorcise-ai/xorcise.git -
agent-orchestra
Datasets & benchmarksAgents can work. They cannot declare success. A governed coordination protocol for multi-agent AI work: eight contracts (task, resource, ownership, handoff, verification, evidence, approval, recovery) enforced in code, plus a reproducible conformance benchmark. Apache-2.0 code, CC BY 4.0 specification, zero dependencies, Node 18+.
git clone https://github.com/3338902669-ops/agent-orchestra.git -
lasagna
Datasets & benchmarksSwap one RAG pipeline stage, hold the rest fixed, and measure whether the gain survives composition. The framework behind "Retrieval Augmented Generation Compositionality": 450 configurations, eight benchmarks, corpora from 5K to 200K documents.
git clone https://github.com/dolev31/lasagna.git
Indexed entries aren't reviewed or tested by InnerVault. Details come straight from each source. Anything that can reach your accounts, files or browser can do damage, so read the source before you install it.