Evaluation & benchmarks for beginner learners
420 resources · page 1 of 14
China's Most Powerful Large Model: Who Is the Real No.1 Among All Claimants? (36Kr)
by 36Kr
36Kr article comparing Chinese models across SuperCLUE, Artificial Analysis and Arena rankings.
3D Arena (HF Space)
by 3D Arena
3D Arena Space: voting arena and leaderboard for AI-generated 3D models.
3DGen-Arena (3DTopia)
by 3DTopia
3DTopia 3DGen-Arena Space: arena comparing open-source text- and image-to-3D generation models.
AA-3D
Artificial Analysis AA-3D benchmark comparing how well AI models build interactive 3D experiences, judged by blind panel comparisons.
Artificial Analysis Intelligence Benchmarking methodology
Artificial Analysis methodology page describing the benchmarks, weights and scoring behind its Intelligence Index for LLMs.
ActivityNet
by ActivityNet
ActivityNet GitHub repo: tools and evaluation code for the ActivityNet human activity benchmark.
AfroBench
Masakhane AfroBench collection on Hugging Face: datasets for benchmarking LLMs across many African languages.
AfroBench site/leaderboard
by McGill NLP
AfroBench site: benchmark evaluating LLMs on 64 African languages across 15 NLP tasks.
AgentBench
by THUKEG
AgentBench GitHub repo (Tsinghua): benchmark evaluating LLMs as agents across multiple interactive environments.
AgentBoard
HKUST AgentBoard GitHub repo: analytical evaluation board for multi-turn LLM agents with progress tracking.
ethz-spylab/agentdojo
by SPY Lab
ETH SPY Lab AgentDojo GitHub repo: environment evaluating prompt injection attacks and defences for LLM agents.
ai-safety-institute/AgentHarm
UK AISI AgentHarm dataset on Hugging Face: harmful agentic tasks for measuring whether LLM agents comply with misuse requests.
AGI-Eval (agi-eval.cn)
AGI-Eval: Chinese arena-style platform for evaluating large language models.
AI Fairness 360
by Trusted-AI
AIF360 GitHub repo (originally IBM): fairness metrics and bias mitigation algorithms for datasets and models.
AI Verify Foundation / Project Moonshot (Singapore)
Singapore's AI Verify Foundation site, home of the AI Verify testing toolkit and Project Moonshot for LLM red-teaming and benchmarking.
AI4Bharat Launches Indic LLM Arena (CXO Digital Pulse)
CXO Digital Pulse report on AI4Bharat's Indic LLM Arena, a crowd-sourced evaluation platform for Indian languages.
Aider polyglot leaderboard
by Aider
Aider LLM leaderboards: model scores on polyglot code-editing exercises with cost.
AILuminate
by MLCommons
MLCommons AILuminate benchmark that rates the safety of general-purpose chat models' responses to hazardous text prompts.
AIR-Bench 2024
by Stanford Center for Research on Foundation Models
AIR-Bench 2024 GitHub repo from Stanford CRFM: safety benchmark with a risk taxonomy drawn from regulations and company policies.
AlignBench
by THUKEG
AlignBench GitHub repo: multi-dimensional benchmark for evaluating alignment of Chinese LLMs, with reference answers.
AlpaSim E2E Closed-Loop Challenge 2026
by OpenDriveLab
OpenDriveLab AlpaSim 2026 challenge: closed-loop benchmark for autonomous-driving policies on safety, generalisation and efficiency.
AndroidWorld
Google Research AndroidWorld GitHub repo: environment and benchmark for autonomous agents on Android.
Apollo Research
Apollo Research site, an AI safety lab publishing evaluations of scheming and deceptive behaviour in frontier models.
AppWorld
by AppWorld team (Stony Brook University)
AppWorld benchmark: a simulated world of everyday apps and APIs for testing coding agents on stateful multi-app tasks.
ArabicMMLU
ArabicMMLU GitHub repo: Arabic knowledge benchmark built from school exams across North African, Levantine and Gulf regions.
Inception Multilingual Leaderboards
by Inception42
Inception multilingual leaderboards Hugging Face Space, including AraGen for Arabic generation.
ARC-AGI-3
ARC-AGI-3 page from ARC Prize: an interactive benchmark where agents must learn the rules of novel game environments.
ARC-AGI leaderboard
ARC Prize leaderboard: model scores on ARC-AGI abstraction benchmarks plotted against cost per task.
Arcade Learning Environment
Documentation for the Arcade Learning Environment, the Atari 2600 interface used to train and evaluate reinforcement-learning agents.
Arena Raises $200M Series B at $3.1B Valuation
Arena blog post on its Series B and the launch of its AI Alignment Index, measuring unsafe agent behaviour on real traces.
Artificial Analysis
Artificial Analysis: independent benchmarks comparing AI models and API providers on intelligence, price, speed and latency.
Artificial Analysis ASR leaderboard
Artificial Analysis leaderboard for non-streaming speech-to-text, comparing word error rate, processing speed and price per model.