Glossaire des benchmarks
Glossaire des benchmarks
Clear definitions connected directly to the benchmark, domain, and media catalogs used throughout Easy Benchmarks.
- Metrics
- 17
- Domains
- 8
- Media benchmarks
- 5
- Definitions
- 30
AIMEAdvanced math competition benchmark score.Artificial Analysis Coding IndexArtificial Analysis aggregate coding score.Artificial Analysis Intelligence IndexArtificial Analysis aggregate LLM intelligence score.Artificial Analysis Math IndexArtificial Analysis aggregate math score.Blended PricePrice per 1M tokens using the Artificial Analysis 3:1 blended ratio.CodingCoding leaderboard from coding-specific benchmark metrics.GPQAGraduate-level science and reasoning benchmark score.Humanity's Last ExamDifficult broad knowledge and reasoning benchmark score.Image EditingArtificial Analysis Elo ratings for models that edit an input image.Image-to-VideoArtificial Analysis Elo ratings for video generation from an input image.Input PriceInput token price per 1M tokens.LiveBench MathematicsCurrent LiveBench Mathematics category score.LiveCodeBenchCompetitive coding and code-generation benchmark score.MathMath leaderboard from math-specific benchmark metrics.MATH-500Mathematical problem-solving benchmark score.MMLU-ProKnowledge and reasoning benchmark score.Output PriceOutput token price per 1M tokens.Output SpeedMedian output tokens generated per second.OverallGeneral leaderboard using the Artificial Analysis Intelligence Index.Price & ValueCost leaderboard using input, output, blended price, and value index.Reasoning & KnowledgeReasoning leaderboard across GPQA, HLE, MMLU-Pro, and related metrics.SciCodeScientific coding benchmark score.SpeedRuntime leaderboard using output speed and latency.Text-to-ImageArtificial Analysis Elo ratings for image generation from text prompts.Text-to-SpeechArtificial Analysis Elo ratings for text-to-speech model quality.Text-to-VideoArtificial Analysis Elo ratings for video generation from text prompts.Time to First TokenMedian seconds until the first output token.TranslationNo direct translation benchmark is included in the current Artificial Analysis LLM endpoint.Value IndexOverall intelligence divided by blended price. Higher means more benchmark score per dollar.WritingNo direct writing benchmark is included in the current Artificial Analysis LLM endpoint.