Rúbrica
Aprobado o suspenso binario por comprobación
¿Siguió el modelo las instrucciones de la tarea, identificó requisitos ocultos en los archivos de origen, utilizó las pruebas correctas y llegó a las conclusiones adecuadas?
AA-Briefcase v1.1: Agentic Knowledge Work Benchmark
Índice de InteligenciaConjunto de datos privadoAgénticoTrabajo del conocimientoUso de herramientasRedacciónRazonamientoSeguimiento de instruccionesNegocios
A 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.
AA-Briefcase evalúa modelos en cuatro proyectos de trabajo del conocimiento de varias semanas, con miles de archivos de entrada y 91 tareas en total. En los distintos escenarios, los modelos deben completar flujos de trabajo profesionales realistas en ámbitos como ciencia de datos, gestión de productos y estrategia corporativa. Cada escenario es un flujo de trabajo de varias semanas que el agente recorre en orden, con varias tareas por semana. Cada tarea produce un entregable evaluado mediante una rúbrica de comprobaciones. Aunque las tareas de un escenario comparten archivos y contexto entre semanas, actualmente los modelos completan cada tarea en una ejecución independiente, sin conservar sus entregas anteriores.
Cada tarea se evalúa con tres tipos de comprobaciones:
Calidad analítica
Comparación por pares
Frente a la entrega de otro modelo, ¿qué entregable es más completo, riguroso desde el punto de vista analítico y está mejor fundamentado?
Presentación
Comparación por pares
Frente a la entrega de otro modelo, ¿cuál tiene una presentación más profesional?
Se ha publicado un quinto escenario a través de Hugging Face como muestra de la estructura, las entregas y la evaluación del escenario. Es solo una demostración y no cuenta para los resultados oficiales de AA-Briefcase.
Resultados
Elo de AA-Briefcase
AA-Briefcase v1.1 is an agentic knowledge work benchmark developed by Artificial Analysis. AA-Briefcase Elo is a combined metric that aggregates rubric pass rate, analytical quality Elo and presentation Elo · Higher is better
Not publicly available
Elo de AA-Briefcase vs. costo por tarea
Elo de AA-Briefcase · Costo por tarea (USD)
Most attractive quadrant
Pareto line
Costo
Costo por tarea de AA-Briefcase
Costo promedio (USD) por tarea de ejecutar AA-Briefcase, calculado a partir del uso de tokens y los precios de los modelos, con tasas representativas de aciertos de caché
Not publicly available
Ejemplo de tarea, entregas y evaluación
Explora una semana representativa de AA-Briefcase del escenario público de diligencia debida disponible en Hugging Face. Los resultados y la evaluación mostrados ilustran qué evalúa AA-Briefcase. Las puntuaciones corresponden a un conjunto representativo de modelos. Las entregas y los veredictos de este escenario representativo no contribuyen al Elo de AA-Briefcase ni a otras puntuaciones del benchmark.
Comparaciones de puntuaciones
Elo de AA-Briefcase vs. Artificial Analysis Intelligence Index
Elo de AA-Briefcase · Artificial Analysis Intelligence Index
Most attractive quadrant
Resultados por tipo de archivo
Rendimiento en AA-Briefcase desglosado por tipo de archivo del entregable (Excel, PowerPoint, PDF, Word u otro).
Tasa de aprobación de la rúbrica de AA-Briefcase por tipo de archivo (normalizada)
Tasa de aprobación de la rúbrica según el tipo de archivo del entregable · Las puntuaciones se normalizan por tipo de archivo entre todos los modelos evaluados: el verde representa la puntuación más alta para ese tipo de archivo y el rojo, la más baja
Uso de tokens
Tokens de salida por tarea de AA-Briefcase
Promedio de tokens de razonamiento y respuesta consumidos por tarea de AA-Briefcase
Not publicly available
Velocidad
Tiempo por tarea
Wall-clock time (minutes) per task: answer and reasoning generation plus tool execution time · Lower is better
Turnos
Promedio de turnos por tarea
Average number of model turns per AA-Briefcase task · Lower is better
Not publicly available
Uso de herramientas
Invocaciones de herramientas de cada agente durante AA-Briefcase: recuentos por categoría, promedio de llamadas por turno y cobertura de exploración de las fuentes.
Desglose de llamadas a herramientas de AA-Briefcase, promedio por tarea
Promedio de invocaciones de herramientas por tarea de AA-Briefcase, agrupadas por intención
Not publicly available
Tamaño del modelo (solo modelos de pesos abiertos)
Elo de AA-Briefcase vs. parámetros totales
Elo de AA-Briefcase · Tamaño en parámetros (miles de millones) · Solo modelos de pesos abiertos
Most attractive quadrant
Puntuación vs. fecha de lanzamiento
Elo de AA-Briefcase vs. fecha de lanzamiento
Elo de AA-Briefcase · Fecha de lanzamiento del modelo
Most attractive region
Clasificación
Creador | Nombre | Elo | IC | Fecha de lanzamiento | |
|---|---|---|---|---|---|
| 1 |  Anthropic | Claude Sonnet 5.5 (Max, Default Fallback) | 1824 | -12 / +13 | sept 2026 |
| 2 | Anthropic | Claude Opus 5.5 (Max, Default Fallback) | 1808 | -13 / +12 | sept 2026 |
| 3 | Anthropic | Claude Opus 5.5 (Xhigh, Default Fallback) | 1768 | -11 / +11 | sept 2026 |
| 4 | Anthropic | Claude Sonnet 5.5 (Xhigh, Default Fallback) | 1751 | -14 / +13 | sept 2026 |
| 5 | Anthropic | Claude Opus 5.5 (High, Default Fallback) | 1690 | -10 / +10 | sept 2026 |
| 6 | Anthropic | Claude Fable 5.1 (Max, Default Fallback) | 1676 | -10 / +10 | sept 2026 |
| 7 | Anthropic | Claude Opus 5 (Max) | 1662 | -8 / +9 | jul 2026 |
| 8 | Anthropic | Claude Fable 5.1 (Xhigh, Default Fallback) | 1657 | -8 / +9 | sept 2026 |
| 9 |  SpaceXAI | Grok 4.7 (Xhigh) | 1645 | -10 / +11 | sept 2026 |
| 10 | Anthropic | Claude Sonnet 5.5 (High, Default Fallback) | 1640 | -9 / +11 | sept 2026 |
| 11 | Anthropic | Claude Opus 5 (Xhigh) | 1636 | -8 / +9 | jul 2026 |
| 12 | SpaceXAI | Grok 4.7 (High) | 1633 | -10 / +11 | sept 2026 |
| 13 | Anthropic | Claude Opus 5.5 (Medium, Default Fallback) | 1628 | -10 / +11 | sept 2026 |
| 14 |  Alibaba | Qwen3.8 Max (0902) | 1621 | -10 / +10 | sept 2026 |
| 15 |  Meta | Muse Spark 1.3 (Max) | 1583 | -9 / +10 | sept 2026 |
Explorar evaluaciones
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.
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 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. The agent reads case documents in a sandbox and produces legal deliverables (e.g., memos, disclosure schedules, deposition summaries), graded criterion-by-criterion by a single LLM rubric judge.
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.
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.
GPQA Diamond Benchmark Leaderboard
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.