AA-3D: Interactive 3D Experience Benchmark
AA-3D
Ranking do AA-3D
Faixa | Modelo | Elo | |
|---|---|---|---|
| 1 | 1-3 |  Claude Sonnet 5.5 (Max, Default Fallback) | 1194±34Intervalo de 95%: 1160 a 1228 |
| 2 | 1-3 | Claude Opus 5.5 (Max, Default Fallback) | 1190±35Intervalo de 95%: 1155 a 1225 |
| 3 | 1-4 |  GPT-6 Astra (Max) | 1165±34Intervalo de 95%: 1131 a 1199 |
| 4 | 4-6 | Claude Fable 5.1 (Max, Default Fallback) | 1131±34Intervalo de 95%: 1098 a 1165 |
| 5 | 4-6 | GPT-6 Sol (Max) | 1123±34Intervalo de 95%: 1089 a 1156 |
| 6 | 4-6 | GPT-6.1 Sol (Max) | 1114±33Intervalo de 95%: 1081 a 1148 |
| 7 | 7 |  Muse Spark 1.3 (Max) | 1035±32Intervalo de 95%: 1003 a 1067 |
| 8 | 8 | GPT-6 Luna (Max) | 1000ÂncoraFixado em 1000 como ponto de referência da escala Elo deste ranking. |
| 9 | 8-13 | Claude Haiku 5.5 (Max) | 994±32Intervalo de 95%: 962 a 1026 |
| 10 | 8-13 |  Grok 4.7 (Xhigh) | 994±32Intervalo de 95%: 962 a 1026 |
| 11 | 8-13 |  Kimi K3 (Max) | 982±33Intervalo de 95%: 949 a 1015 |
| 12 | 8-14 |  MiMo-V2.6-Pro | 977±32Intervalo de 95%: 945 a 1009 |
| 13 | 9-14 |  Gemini 3.8 Flash (High) | 965±32Intervalo de 95%: 932 a 997 |
| 14 | 12-15 |  Qwen3.8 Max (0902) | 948±32Intervalo de 95%: 916 a 981 |
| 15 | 14-15 |  GLM-5.3-Flash | 931±32Intervalo de 95%: 899 a 963 |
| 16 | 16-17 |  Mistral Large 4 Preview | 868±32Intervalo de 95%: 835 a 900 |
| 17 | 16-17 |  Step 5 Preview | 848±33Intervalo de 95%: 814 a 881 |
| 18 | 18 | GLM-5.3 (Max) | 782±35Intervalo de 95%: 747 a 816 |
| 19 | 19 |  MiniMax-M3 | 739±34Intervalo de 95%: 704 a 773 |
| 20 | 20-21 |  Inkling (Xhigh) | 579±41Intervalo de 95%: 538 a 620 |
| 21 | 20-21 |  Nemotron 3 Ultra 550B A55B (Reasoning) | 569±45Intervalo de 95%: 524 a 614 |
AA-3D Elo
AA-3D Elo
AA-3D: Elo vs. custo por tarefa
Tarefas de exemplo
Escolha um domínio, uma tarefa e um modelo para carregar a cena que esse modelo criou.
Tarefa
Modelo
Create an interactive adult man performing weightlifting exercises, with views that expose his muscles, skeleton and organs in anatomically accurate form.
21 de 21 modelos têm uma cena para esta tarefa na votação cega.
Elo vs. Intelligence Index
Elo do AA-3D vs. Intelligence Index
Taxa de vitórias por domínio
Cada célula mostra a parcela dos votos que um modelo venceu em um domínio de tarefas, com base nos votos do ranking atual. A visão equilibrada por adversário tira a média das taxas de vitórias do modelo contra cada adversário do domínio, para que cada adversário conte igualmente.
Taxa de vitórias do AA-3D por domínio
Custo
AA-3D: custo por tarefa
Uso de tokens
AA-3D: tokens de saída por tarefa
Turnos por cena
Turnos por cena do AA-3D
Uso de ferramentas
Chamadas de ferramentas por cena do AA-3D
Como o AA-3D funciona
Cada modelo trabalha como agente em um sandbox offline. Ele recebe uma descrição escrita de uma cena interativa e um contrato de entrega comum a todas as tarefas. O modelo escreve código, renderiza a cena, analisa as próprias capturas de tela e faz revisões. Ele tem à disposição Blender, física Rapier, Python, um navegador headless e uma biblioteca de recursos CC0. Uma execução termina após 3 horas ou 250 turnos. O resultado é uma página index.html que executa a cena em um navegador.
Por dentro de uma execução do agente
Estas reproduções usam os dados do lançamento do AA-3D. Escolha uma tarefa para reproduzir a execução de cada modelo em um relógio compartilhado. Passe o cursor sobre uma linha para ver cada ação e clique nela para ver para onde foi o tempo.
Veja uma cena sendo criada
Pontuação vs. data de lançamento
AA-3D: pontuação vs. data de lançamento
Perguntas frequentes
O AA-3D v1.0 mede o quão bem os modelos de IA criam experiências 3D interativas. Cada modelo recebe uma descrição escrita da tarefa e cria uma única página web com three.js em um sandbox offline. O benchmark tem 55 tarefas em 7 domínios. Leia a metodologia do AA-3D
Os modelos são classificados com um sistema de pontuação Elo derivado de comparações às cegas feitas pelo nosso painel humano recrutado. Os participantes do painel interagem com duas cenas criadas para a mesma tarefa e escolhem a que preferem, sem saber qual modelo criou cada uma. Pontuações Elo mais altas indicam que um modelo é preferido com mais frequência. Experimente a Votação do AA-3D
Explorar avaliações
Artificial Analysis Intelligence Index v4.3.2A composite benchmark aggregating ten challenging evaluations to provide a holistic measure of AI capabilities across mathematics, science, coding, and reasoning.
Artificial Analysis Openness IndexA composite measure providing an industry standard to communicate model openness for users and developers.
AA-Briefcase v1.1: Agentic Knowledge Work BenchmarkA private evaluation developed by Artificial Analysis for frontier agentic capability in long-horizon knowledge work, testing agents on realistic business workflows that require deliverables such as spreadsheets, presentations, and memos.
GDPval-AA v2.1 LeaderboardGDPval-AA v2.1 is Artificial Analysis' evaluation framework for OpenAI's GDPval dataset. It tests AI models on real-world tasks across 44 occupations and 9 major industries. Models are given shell access and web browsing capabilities in an agentic loop via Stirrup to solve tasks, with Elo ratings derived from blind pairwise comparisons.
AA-3D: Interactive 3D Experience BenchmarkA benchmark developed by Artificial Analysis for how well AI models build interactive 3D experiences. Models work as agents in an offline sandbox via Stirrup to turn a written brief into a three.js scene, with Elo ratings derived from blind pairwise votes.
APEX-Agents-AA Benchmark LeaderboardArtificial Analysis' implementation of the APEX-Agents benchmark, testing AI agents on long-horizon, cross-application tasks in professional-services environments with realistic application tooling.
AA-AnalystAgent Benchmark LeaderboardArtificial Analysis' data analysis benchmark, testing AI agents on their ability to work with spreadsheets and documents to answer quantitative questions a Business Analyst or Data Analyst would face day-to-day.
AutomationBench-AA: Agentic SaaS Workflow BenchmarkA benchmark measuring agentic task completion across simulated SaaS application environments, scoring the share of each task's objectives completed without guardrail violations.
Harvey LAB-AA v1.1 Benchmark LeaderboardArtificial Analysis' implementation of Harvey's Legal Agent Benchmark (LAB), testing AI agents on real-world legal work from Harvey's dataset of 120 private tasks spanning 24 legal practice areas. v1.1, built in collaboration with Harvey, checks every deliverable against the task's source documents for hallucinations, so work that passes every rubric criterion but contains a material hallucination receives no credit. A panel of three judges grades each deliverable criterion-by-criterion.
GDP.pdf Benchmark LeaderboardArtificial Analysis' implementation of Surge AI's GDP.pdf benchmark, testing whether language models can reason over long, real-world professional documents and satisfy detailed task-specific criteria.
EnterpriseOps-Gym-AA Benchmark LeaderboardArtificial Analysis' independent implementation of ServiceNow's EnterpriseOps-Gym, an agentic benchmark testing whether LLM agents can complete stateful, multi-step enterprise workflows across eight business domains via live tool use, graded on the final state of the underlying databases.
Terminal-Bench 4.0 Benchmark LeaderboardA harder 66-task benchmark of complex terminal work across software, machine learning, science, operations, security, hardware, and media, with recalibrated compute and time allowances and improved instructions, environments, and verifiers.
Terminal-Bench-Science 0.1 Benchmark LeaderboardA 70-task benchmark of research workflows authored and reviewed by domain experts across the life, physical, earth, mathematical, and engineering sciences, each completed in a terminal and checked by its own set of tests.
FrontierCode 1.1 Benchmark LeaderboardArtificial Analysis' implementation of Cognition's FrontierCode 1.1, a private agentic coding benchmark that tests models on completing coding tasks in a way that would actually get merged by maintainers of real-world repositories.
Artificial Analysis Long Context Reasoning Benchmark LeaderboardA challenging benchmark measuring language models' ability to extract, reason about, and synthesize information from long-form documents ranging from 10k to 100k tokens (measured using the cl100k_base tokenizer).
AA-Omniscience: Knowledge and Hallucination BenchmarkA benchmark measuring factual recall and hallucination across various economically relevant domains.
SciCode Benchmark LeaderboardA scientist-curated coding benchmark featuring 288 test set subproblems from 80 laboratory problems across 16 scientific disciplines.
Humanity's Last Exam Benchmark LeaderboardA frontier-level benchmark with 2,500 expert-vetted questions across mathematics, sciences, and humanities, designed to be the final closed-ended academic evaluation.
CritPt Benchmark LeaderboardA benchmark designed to test LLMs on research-level physics reasoning tasks, featuring 71 composite research challenges.
The most challenging 198 questions from GPQA, where PhD experts achieve 65% accuracy but skilled non-experts only reach 34% despite web access.
ITBench-AA Benchmark LeaderboardArtificial Analysis' implementation of IBM's ITBench benchmark, testing AI agents on Kubernetes incident root-cause analysis from offline incident snapshots. The agent inspects alerts, events, traces, and topology and identifies the contributing-factor entities (deployments, pods, namespaces, network policies, etc.) responsible for the failure.
MMMU-Pro Benchmark LeaderboardAn enhanced MMMU benchmark that eliminates shortcuts and guessing strategies to more rigorously test multimodal models across 30 academic disciplines.
IFBench Benchmark LeaderboardA benchmark evaluating precise instruction-following generalization on 58 diverse, verifiable out-of-domain constraints that test models' ability to follow specific output requirements.
Medical Long Context Reasoning (MLCR-AA)An open benchmark from Wisedocs measuring how well models reason over long, fragmented medical records, performing the multi-document synthesis claims professionals rely on when reviewing insurance and healthcare cases.
𝜏³-Banking Benchmark LeaderboardA fintech customer-support benchmark from the 𝜏-Knowledge framework that tests whether agents can navigate a large unstructured knowledge base and execute multi-step tool calls to resolve realistic banking workflows.
Terminal-Bench 2.1 Benchmark LeaderboardA verified refresh of Terminal-Bench 2.0 — 89 curated tasks across software engineering, system administration, data processing, model training, and security, with environment and instruction fixes so scores reflect agent capability rather than environment gaps.
Terminal-Bench Hard Benchmark LeaderboardAn agentic benchmark evaluating AI capabilities in terminal environments through software engineering, system administration, and data processing tasks.
𝜏²-Bench Telecom Benchmark LeaderboardA dual-control conversational AI benchmark simulating technical support scenarios where both agent and user must coordinate actions to resolve telecom service issues.
MMLU-Pro Benchmark LeaderboardAn enhanced version of MMLU with 12,000 graduate-level questions across 14 subject areas, featuring ten answer options and deeper reasoning requirements.
LiveCodeBench Benchmark LeaderboardA contamination-free coding benchmark that continuously harvests fresh competitive programming problems from LeetCode, AtCoder, and CodeForces, evaluating code generation, self-repair, and execution.
MATH-500 Benchmark LeaderboardA 500-problem subset from the MATH dataset, featuring competition-level mathematics across six domains including algebra, geometry, and number theory.
AIME 2025 Benchmark LeaderboardAll 30 problems from the 2025 American Invitational Mathematics Examination, testing olympiad-level mathematical reasoning with integer answers from 000-999.
Global-MMLU-Lite Benchmark LeaderboardA lightweight, multilingual version of MMLU, designed to evaluate knowledge and reasoning skills across a diverse range of languages and cultural contexts.