Benchmark Catalog

Browse by category

34478 public benchmarks carry scores in the current fit, with source links and saturation forecasts. The catalogue on this page tracks more boards than the fit uses. The most broadly covered benchmarks are listed below.

BenchmarkCategoryMetricModelsSaturation forecast
Open Portuguese LLM - ASSIN2 RTEMacro F1 (%)638Estimated already saturated.
Open Portuguese LLM - BLUEXAccuracy (%)638Estimated already saturated.
Open Portuguese LLM - ENEMAccuracy (%)638Estimated already saturated.
Open Portuguese LLM - FaQuAD NLIMacro F1 (%)638Estimated already saturated.
Open Portuguese LLM - OAB ExamsAccuracy (%)638Around February 2027.
Open LLM Leaderboard - BBHScore604Estimated already saturated.
Open LLM Leaderboard - GPQAScore604Estimated already saturated.
Open LLM Leaderboard - IFEvalScore604Estimated already saturated.
Open LLM Leaderboard - MATH Level 5Score604Estimated already saturated.
Open LLM Leaderboard - MMLU-ProScore604Estimated already saturated.
Open LLM Leaderboard - MuSRScore604Estimated already saturated.
AA Humanity's Last ExamGeneral Knowledge & LanguageAccuracy (%)546Around April 2027.
AA CritPtMath, Logic & ReasoningAccuracy (%)498Around May 2027.
AA GPQA DiamondMath, Logic & ReasoningAccuracy (%)495Estimated already saturated.
AA Long Context ReasoningGeneral Knowledge & LanguageAccuracy (%)484Around February 2027.
AA OmniscienceGeneral Knowledge & LanguageScore479Around December 2027.
AA-Omniscience AccuracyAccuracy (%)479Around December 2027.
AA-Omniscience Hallucination RateHallucination Rate (%)479Not forecast.
Wolfram LLM Benchmarking ProjectCoding & Software EngineeringCorrect Functionality (%)381Around February 2027.
AA IFBenchGeneral Knowledge & LanguageAccuracy (%)361Around February 2027.
AA TAU-2 BenchAgentic & Tool UseAccuracy (%)351Estimated already saturated.
AA Terminal-Bench HardCoding & Software EngineeringAccuracy (%)345Around February 2027.
LLM Stats ScoreLLM Stats Score (conservative rating)342Around December 2027.
Tau2-Bench TelecomAgentic & Tool UseSuccess Rate (%)340Estimated already saturated.
Open LLM Leaderboard v1 - ARC ChallengeNormalized accuracy (%) (25-shot)333Estimated already saturated.
Open LLM Leaderboard v1 - GSM8KAccuracy (%) (5-shot)333Estimated already saturated.
Open LLM Leaderboard v1 - HellaSwagNormalized accuracy (%) (10-shot)333Estimated already saturated.
Open LLM Leaderboard v1 - MMLUAccuracy (%) (5-shot)333Estimated already saturated.
Open LLM Leaderboard v1 - TruthfulQA MC2MC2 (%) (0-shot)333Estimated already saturated.
Open LLM Leaderboard v1 - WinoGrandeAccuracy (%) (5-shot)333Estimated already saturated.
Tinybird AI SQL Benchmark - First-Attempt Success RateQuestions answered with a valid query on the first attempt (%)329Estimated already saturated.
Tinybird AI SQL Benchmark - Success RateQuestions answered with a valid query within 3 retries (%)329Estimated already saturated.
Chatbot Arena (Text - English)Arena Score328Forecast unavailable.
Chatbot Arena (Text - Hard Prompts)Arena Score328Forecast unavailable.
Chatbot Arena (Text - Instruction Following)Arena Score328Forecast unavailable.
Chatbot Arena (Text - Non-English)Arena Score328Forecast unavailable.
Chatbot Arena (Text - Software & IT Services)Arena Score328Forecast unavailable.
Chatbot Arena (Text)Elo328Around November 2027.
Chatbot Arena (Text - Creative Writing)Arena Score327Forecast unavailable.
Chatbot Arena (Text - Entertainment, Sports & Media)Arena Score327Forecast unavailable.
Chatbot Arena (Text - Hard Prompts (English))Arena Score327Forecast unavailable.
Chatbot Arena (Text - Life, Physical & Social Science)Arena Score327Forecast unavailable.
Chatbot Arena (Text - Multi-Turn)Arena Score327Forecast unavailable.
Chatbot Arena (Text - Writing, Literature & Language)Arena Score327Forecast unavailable.
Chatbot Arena (Text - Coding)Arena Score326Forecast unavailable.
AI Chess Leaderboard (Reasoning)Games & PuzzlesElo325Around January 2028.
Chatbot Arena (Text - Business, Management & Financial Ops)Arena Score324Forecast unavailable.
Tinybird AI SQL Benchmark - ExactnessResult exactness vs human reference queries (0-100)323Around July 2028.
SpeechMap Compliance% Requests Completed321Forecast unavailable.
ObviousBenchAnswer pass³ (%)321Estimated already saturated.
Chatbot Arena (Text - Math)Arena Score320Forecast unavailable.
Chatbot Arena (Text - Longer Query)Arena Score317Forecast unavailable.
Chatbot Arena (Text - Mathematical)Arena Score317Forecast unavailable.
Chatbot Arena (Text - Chinese)Arena Score312Forecast unavailable.
Chatbot Arena (Text - Legal & Government)Arena Score312Forecast unavailable.
Chatbot Arena (Text - Medicine & Healthcare)Arena Score312Forecast unavailable.
Chatbot Arena (Text - Russian)Arena Score310Forecast unavailable.
Open Korean LLM - Ko-EQ-BenchEQ-Bench score308Estimated already saturated.
Open Korean LLM - Ko-GPQA-DiamondNormalized accuracy (%)308Estimated already saturated.
Open Korean LLM - Ko-GSM8K (flexible extract)Exact match, flexible extract (%)308Estimated already saturated.

Interactive version: theaggregate.ai/benchmarks/catalog · How It Works · Data refreshed daily, snapshot 2026-11-14.