Benchmarks

PORTULAN ExtraGLUE is a natural language processing evaluation suite designed to benchmark language models on Portuguese-language tasks. Developed under the...

The Gauss-AI-Compute Competition is a German challenge launched in 2025 that invites research groups, companies in collaboration with German research institu...

EEE-Bench is an evaluation suite designed to assess the accuracy of AI models in specialized technical domains. It focuses on engineering, energy, and enviro...

MR-Ben is an open-source evaluation suite designed to assess meta-reasoning and system-2 thinking capabilities in large language models (LLMs). It measures p...

SimpleBench is a multiple-choice text benchmark designed to evaluate large language models (LLMs) on reasoning, social intelligence, and linguistic adversari...

EffiBench is a benchmark leaderboard hosted on Hugging Face that evaluates the efficiency of AI models. It measures two key performance metrics: execution ti...

BhashaBench V1 is a bilingual evaluation benchmark designed to assess large language models on domain-specific knowledge in English and Hindi. It contains 74...

The Indic LLM Leaderboard is a multilingual evaluation benchmark created by CognitiveLab in 2024 to assess large language model performance across English an...

Indic LLM Arena is a crowd-sourced, human-in-the-loop leaderboard developed by AI4Bharat at IIT Madras, launched in 2025 to evaluate large language models on...

VMLU is a Vietnamese-language benchmark designed to evaluate the general knowledge and reasoning capabilities of large language models across 10,880 multiple...

Geekbench AI is a cross-platform benchmark developed by Primate Labs that measures AI inference performance on mobile and tablet devices running Android and...

AI BENCHY Leaderboard is a multilingual benchmark platform that evaluates and ranks AI language models across eight test categories, including General Intell...

GermanBenchmark is an evaluation suite consisting of four German-language datasets translated from established English benchmarks: ARC-Challenge, HellaSwag,...

SuperGLEBer is a German-language evaluation suite designed to benchmark the natural language understanding capabilities of language models across 29 tasks. I...

COLE (Comprehensive benchmark for French Language Understanding Evaluation) is an evaluation suite designed to assess the French language understanding capab...

Compar:IA is an open French-language platform developed by the French Ministry of Culture's digital service in partnership with PEReN (Pôle d'Expertise de la...

Danoliterate is a benchmark and evaluation suite developed by DTU Compute for assessing the natural language generation capabilities of large language models...

The Benchmark Project is a Danish-language evaluation suite developed by the Centre for Language Technology at the University of Copenhagen, running from 202...

The Tucano 2 Evaluation Suite is a benchmark collection designed to assess language model performance specifically for Portuguese, introduced alongside the T...

The European Portuguese Benchmark is a language model evaluation suite released in 2025 as part of the broader EuroEval framework, designed to assess model p...

BRFauna Gen-Eval is a benchmark evaluation suite designed to assess the performance of AI language models on questions related to Brazilian fauna, covering t...

PoETa v2 is a Portuguese-language benchmark suite designed to evaluate large language models across 44 tasks, comprising 12 natively created tasks and 32 tra...

Alvorada-Bench is a Portuguese-language benchmark designed to evaluate the performance of language models on Brazilian educational and general knowledge task...

The Portuguese Language AI Models Benchmark is a leaderboard created by Artificial Analysis that evaluates and compares the performance of large language mod...

CALAME-PT (Context-Aware LAnguage Modeling Evaluation for Portuguese) is a zero-shot language modeling benchmark designed to evaluate the capabilities of Por...

BGPA (Brazilian Graduate Proficiency in Accounting) is a text-based benchmark designed to evaluate the accounting and finance knowledge of large language mod...

BLUEX is a multimodal benchmark designed to evaluate the performance of AI language models on Brazilian university entrance exam questions. It consists of 1,...

Napolab is a multilingual evaluation suite released in 2023 for benchmarking natural language processing models across Portuguese, Catalan, English, Galician...

lm-evaluation-harness-pt is a Portuguese-language evaluation suite forked from EleutherAI's lm-evaluation-harness, developed by eduagarcia at the Center of E...

The Open Portuguese LLM Leaderboard is a benchmark and evaluation platform hosted on Hugging Face that ranks large language models based on their performance...

TangoIA Benchmarks is a Spanish-language evaluation leaderboard developed in Argentina for assessing the performance of large language models across a range...

The Índice Latinoamericano de Inteligencia Artificial (ILIA) is a regional benchmark and leaderboard that evaluates and ranks 19 Latin American countries acc...

CAMEL-Bench is a comprehensive Arabic multimodal evaluation benchmark comprising 29,036 questions spanning 8 domains and 38 sub-domains, published in 2024 an...

LAraBench is a comprehensive benchmarking suite designed to evaluate large language and speech models on Arabic natural language processing and speech tasks....

The Open Arabic LLM Leaderboard (OALL) is a benchmarking platform developed by the Technology Innovation Institute (UAE) in 2024 for evaluating and comparing...

AI Energy Score is a benchmark and leaderboard hosted on Hugging Face that measures the energy efficiency of AI language models during inference. It evaluate...

MCP-BENCH is a benchmark evaluation suite designed to assess the capabilities of AI agents that interact with Model Context Protocol (MCP) servers. It covers...

ENIGMAEVAL is a benchmark developed by Scale AI, published in 2025, designed to evaluate the reasoning and puzzle-solving capabilities of AI systems. It cons...

SciCode is a benchmark designed to evaluate the ability of AI models to solve scientific research-level coding problems across multiple disciplines, includin...

The ARC Prize Leaderboard is a public benchmarking resource maintained by the ARC Prize Foundation that tracks AI system performance on the ARC-AGI-1 and ARC...

CyberBench is a benchmark evaluation suite designed to assess the cybersecurity capabilities and potential risks of large language models, published at ICLR...

LiveBench is a text-based AI benchmark designed to evaluate large language models across a range of general reasoning and knowledge tasks. It operates as a c...

Humanity's Last Exam is a challenging benchmark developed by the Center for AI Safety and Scale AI in 2025, designed to evaluate the capabilities of advanced...

MASK (Model Alignment between Statements and Knowledge) is a benchmark developed by Scale AI in partnership with the Center for AI Safety (CAIS) in 2025 to e...

SEAL LLM Leaderboards is a benchmarking initiative maintained by Scale AI that evaluates and ranks large language models across a range of capabilities. The...

LiveCodeBench is a contamination-free benchmark designed to evaluate large language models on coding tasks using problems sourced from LeetCode, AtCoder, and...

Aider LLM Leaderboards is a benchmark designed to evaluate large language models on their ability to write and edit code without human intervention. It uses...

SWE-bench is a benchmark developed by Princeton University and Stanford University for evaluating the ability of AI systems to resolve real-world software en...