{"source":"Quelle IA, le classement des modèles de langage","page":"https://quelleia.com/benchmarks/frontiermath_tier_4/","edition":"2026-W40","date":"2026-09-28","echelle":"score_publie est le score de la source, dans son unité. score_0_1 est ce score ramené entre 0 et 1 pour le calcul. niveau_suggere est le score IA que cette seule mesure indique : 100 pour le meilleur modèle au lancement du site, 50 pour GPT-4o de novembre 2024. difficulte est le score IA d'un modèle qui obtient la moitié (score_0_1 = 0,5), pente la raideur de la courbe.","licence":"Réutilisation libre en citant « Quelle IA, quelleia.com » et la date de l'édition. Les mesures renvoient à leurs sources d'origine.","benchmark":{"id":"frontiermath_tier_4","nom":"FrontierMath, palier 4 (version 2025)","tache":"maths","etat":"actif","echelle":"unite","modeles":55,"poids_general":1.25,"difficulte":90.1,"pente":0.1906,"en_une_phrase":"Une cinquantaine de problèmes de niveau recherche, les plus durs de FrontierMath, qui peuvent occuper un mathématicien plusieurs jours.","notation":"Un point par réponse exacte, vérifiée par programme ; le score est la part de problèmes résolus sur le jeu privé.","limites":"Version remplacée en juin 2026 après correction de 12 énoncés et retrait de 7 autres ; Epoch estime le score maximal à 60 %. OpenAI a financé le jeu et en détient une partie.","mainteneur":"Epoch AI","url":"https://epoch.ai/benchmarks/frontiermath","verifie":true},"tache":{"id":"maths","nom":"Mathématiques","poids_general":10,"benchmarks_actifs":8},"modeles_mesures":55,"mesures":72,"classement":[{"rang":1,"modele":"ai-co-mathematician","nom":"AI Co-Mathematician","editeur":"Google DeepMind","page":"https://quelleia.com/modeles/ai-co-mathematician/","score_publie":0.479,"score_0_1":0.7983,"niveau_suggere":97.3,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2026-05-08","source":"https://epoch.ai/benchmarks"},{"rang":2,"modele":"gpt-5-5-pro","nom":"GPT-5.5 Pro","editeur":"OpenAI","page":"https://quelleia.com/modeles/gpt-5-5-pro/","score_publie":0.396,"score_0_1":0.66,"niveau_suggere":93.6,"reflexion":"maximale","harnais":null,"configurations_mesurees":2,"date":"2026-04-23","source":"https://epoch.ai/benchmarks"},{"rang":3,"modele":"gpt-5-4-pro","nom":"GPT-5.4 Pro","editeur":"OpenAI","page":"https://quelleia.com/modeles/gpt-5-4-pro/","score_publie":0.375,"score_0_1":0.625,"niveau_suggere":92.8,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2026-03-06","source":"https://epoch.ai/benchmarks"},{"rang":4,"modele":"gpt-5-5","nom":"GPT-5.5","editeur":"OpenAI","page":"https://quelleia.com/modeles/gpt-5-5/","score_publie":0.354,"score_0_1":0.59,"niveau_suggere":92,"reflexion":"maximale","harnais":null,"configurations_mesurees":1,"date":"2026-04-23","source":"https://epoch.ai/benchmarks"},{"rang":5,"modele":"gpt-5-2-pro","nom":"GPT-5.2 Pro","editeur":"OpenAI","page":"https://quelleia.com/modeles/gpt-5-2-pro/","score_publie":0.313,"score_0_1":0.5217,"niveau_suggere":90.6,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2025-12-24","source":"https://epoch.ai/benchmarks"},{"rang":6,"modele":"claude-opus-4-8","nom":"Claude Opus 4.8","editeur":"Anthropic","page":"https://quelleia.com/modeles/claude-opus-4-8/","score_publie":0.3125,"score_0_1":0.5208,"niveau_suggere":90.6,"reflexion":"maximale","harnais":null,"configurations_mesurees":1,"date":"2026-06-08","source":"https://epoch.ai/benchmarks"},{"rang":7,"modele":"gpt-5-4","nom":"GPT-5.4","editeur":"OpenAI","page":"https://quelleia.com/modeles/gpt-5-4/","score_publie":0.271,"score_0_1":0.4517,"niveau_suggere":89.1,"reflexion":"maximale","harnais":null,"configurations_mesurees":1,"date":"2026-03-06","source":"https://epoch.ai/benchmarks"},{"rang":8,"modele":"claude-opus-4-7","nom":"Claude Opus 4.7","editeur":"Anthropic","page":"https://quelleia.com/modeles/claude-opus-4-7/","score_publie":0.2291666666666666,"score_0_1":0.3819,"niveau_suggere":87.6,"reflexion":"maximale","harnais":null,"configurations_mesurees":1,"date":"2026-04-17","source":"https://epoch.ai/benchmarks"},{"rang":9,"modele":"claude-opus-4-6","nom":"Claude Opus 4.6","editeur":"Anthropic","page":"https://quelleia.com/modeles/claude-opus-4-6/","score_publie":0.229,"score_0_1":0.3817,"niveau_suggere":87.6,"reflexion":"maximale","harnais":null,"configurations_mesurees":4,"date":"2026-02-12","source":"https://epoch.ai/benchmarks"},{"rang":10,"modele":"gpt-5-2","nom":"GPT-5.2","editeur":"OpenAI","page":"https://quelleia.com/modeles/gpt-5-2/","score_publie":0.188,"score_0_1":0.3133,"niveau_suggere":86,"reflexion":"elevee","harnais":null,"configurations_mesurees":4,"date":"2025-12-11","source":"https://epoch.ai/benchmarks"},{"rang":11,"modele":"gemini-3-pro","nom":"Gemini 3 Pro","editeur":"Google DeepMind","page":"https://quelleia.com/modeles/gemini-3-pro/","score_publie":0.1875,"score_0_1":0.3125,"niveau_suggere":86,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2025-11-21","source":"https://epoch.ai/benchmarks"},{"rang":12,"modele":"gemini-3-1-pro","nom":"Gemini 3.1 Pro","editeur":"Google DeepMind","page":"https://quelleia.com/modeles/gemini-3-1-pro/","score_publie":0.167,"score_0_1":0.2783,"niveau_suggere":85.1,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2026-02-19","source":"https://epoch.ai/benchmarks"},{"rang":13,"modele":"muse-spark","nom":"Muse Spark","editeur":"Meta AI","page":"https://quelleia.com/modeles/muse-spark/","score_publie":0.146,"score_0_1":0.2433,"niveau_suggere":84.2,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2026-04-08","source":"https://epoch.ai/benchmarks"},{"rang":13,"modele":"gpt-5-pro","nom":"GPT-5 Pro","editeur":"OpenAI","page":"https://quelleia.com/modeles/gpt-5-pro/","score_publie":0.146,"score_0_1":0.2433,"niveau_suggere":84.2,"reflexion":"elevee","harnais":null,"configurations_mesurees":1,"date":"2025-10-09","source":"https://epoch.ai/benchmarks"},{"rang":15,"modele":"gemini-3-5-flash","nom":"Gemini 3.5 Flash","editeur":"Google DeepMind","page":"https://quelleia.com/modeles/gemini-3-5-flash/","score_publie":0.1458333333333333,"score_0_1":0.2431,"niveau_suggere":84.2,"reflexion":"elevee","harnais":null,"configurations_mesurees":1,"date":"2026-05-25","source":"https://epoch.ai/benchmarks"},{"rang":16,"modele":"kimi-k2-6","nom":"Kimi K2.6","editeur":"Moonshot","page":"https://quelleia.com/modeles/kimi-k2-6/","score_publie":0.1458,"score_0_1":0.243,"niveau_suggere":84.2,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2026-05-08","source":"https://epoch.ai/benchmarks"},{"rang":17,"modele":"glm-5-1","nom":"GLM-5.1","editeur":"Z.ai (Zhipu AI)","page":"https://quelleia.com/modeles/glm-5-1/","score_publie":0.125,"score_0_1":0.2083,"niveau_suggere":83.1,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2026-05-12","source":"https://epoch.ai/benchmarks"},{"rang":17,"modele":"gpt-5","nom":"GPT-5","editeur":"OpenAI","page":"https://quelleia.com/modeles/gpt-5/","score_publie":0.125,"score_0_1":0.2083,"niveau_suggere":83.1,"reflexion":"elevee","harnais":null,"configurations_mesurees":2,"date":"2025-10-30","source":"https://epoch.ai/benchmarks"},{"rang":17,"modele":"gpt-5-1","nom":"GPT-5.1","editeur":"OpenAI","page":"https://quelleia.com/modeles/gpt-5-1/","score_publie":0.125,"score_0_1":0.2083,"niveau_suggere":83.1,"reflexion":"elevee","harnais":null,"configurations_mesurees":2,"date":"2025-11-13","source":"https://epoch.ai/benchmarks"},{"rang":20,"modele":"gemini-2-5-deep-think","nom":"Gemini 2.5 Deep Think","editeur":"Google DeepMind","page":"https://quelleia.com/modeles/gemini-2-5-deep-think/","score_publie":0.104,"score_0_1":0.1733,"niveau_suggere":81.9,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":21,"modele":"qwen-3-6-plus","nom":"Qwen 3.6 Plus","editeur":"Alibaba","page":"https://quelleia.com/modeles/qwen-3-6-plus/","score_publie":0.0833333333333333,"score_0_1":0.1389,"niveau_suggere":80.5,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2026-05-12","source":"https://epoch.ai/benchmarks"},{"rang":22,"modele":"claude-sonnet-4-6","nom":"Claude Sonnet 4.6","editeur":"Anthropic","page":"https://quelleia.com/modeles/claude-sonnet-4-6/","score_publie":0.083,"score_0_1":0.1383,"niveau_suggere":80.5,"reflexion":"budget","harnais":null,"configurations_mesurees":1,"date":"2026-02-21","source":"https://epoch.ai/benchmarks"},{"rang":23,"modele":"gpt-5-4-nano","nom":"GPT-5.4 Nano","editeur":"OpenAI","page":"https://quelleia.com/modeles/gpt-5-4-nano/","score_publie":0.0625,"score_0_1":0.1042,"niveau_suggere":78.8,"reflexion":"elevee","harnais":null,"configurations_mesurees":1,"date":"2026-04-15","source":"https://epoch.ai/benchmarks"},{"rang":23,"modele":"gpt-5-mini","nom":"GPT-5 mini","editeur":"OpenAI","page":"https://quelleia.com/modeles/gpt-5-mini/","score_publie":0.0625,"score_0_1":0.1042,"niveau_suggere":78.8,"reflexion":"elevee","harnais":null,"configurations_mesurees":2,"date":"2025-10-30","source":"https://epoch.ai/benchmarks"},{"rang":23,"modele":"o4-mini","nom":"o4-mini","editeur":"OpenAI","page":"https://quelleia.com/modeles/o4-mini/","score_publie":0.0625,"score_0_1":0.1042,"niveau_suggere":78.8,"reflexion":"elevee","harnais":null,"configurations_mesurees":2,"date":"2025-07-01","source":"https://epoch.ai/benchmarks"},{"rang":26,"modele":"kimi-k2-5","nom":"Kimi K2.5","editeur":"Moonshot","page":"https://quelleia.com/modeles/kimi-k2-5/","score_publie":0.042,"score_0_1":0.07,"niveau_suggere":76.5,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2026-02-02","source":"https://epoch.ai/benchmarks"},{"rang":27,"modele":"gemini-3-flash","nom":"Gemini 3 Flash","editeur":"Google DeepMind","page":"https://quelleia.com/modeles/gemini-3-flash/","score_publie":0.0416666666666666,"score_0_1":0.0694,"niveau_suggere":76.5,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2025-12-17","source":"https://epoch.ai/benchmarks"},{"rang":27,"modele":"claude-opus-4-5","nom":"Claude Opus 4.5","editeur":"Anthropic","page":"https://quelleia.com/modeles/claude-opus-4-5/","score_publie":0.0416666666666666,"score_0_1":0.0694,"niveau_suggere":76.5,"reflexion":"inconnue","harnais":null,"configurations_mesurees":3,"date":"2025-11-25","source":"https://epoch.ai/benchmarks"},{"rang":27,"modele":"qwen-3-6-max-preview","nom":"Qwen 3.6 Max (Preview)","editeur":"Alibaba","page":"https://quelleia.com/modeles/qwen-3-6-max-preview/","score_publie":0.0416666666666666,"score_0_1":0.0694,"niveau_suggere":76.5,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2026-05-28","source":"https://epoch.ai/benchmarks"},{"rang":27,"modele":"claude-sonnet-4-5","nom":"Claude Sonnet 4.5","editeur":"Anthropic","page":"https://quelleia.com/modeles/claude-sonnet-4-5/","score_publie":0.0416666666666666,"score_0_1":0.0694,"niveau_suggere":76.5,"reflexion":"budget","harnais":null,"configurations_mesurees":2,"date":"2025-10-22","source":"https://epoch.ai/benchmarks"},{"rang":27,"modele":"claude-opus-4-1","nom":"Claude Opus 4.1","editeur":"Anthropic","page":"https://quelleia.com/modeles/claude-opus-4-1/","score_publie":0.0416666666666666,"score_0_1":0.0694,"niveau_suggere":76.5,"reflexion":"budget","harnais":null,"configurations_mesurees":1,"date":"2025-08-05","source":"https://epoch.ai/benchmarks"},{"rang":27,"modele":"gemini-2-5-pro-jun-2025","nom":"Gemini 2.5 Pro (Jun 2025)","editeur":"Google DeepMind","page":"https://quelleia.com/modeles/gemini-2-5-pro-jun-2025/","score_publie":0.0416666666666666,"score_0_1":0.0694,"niveau_suggere":76.5,"reflexion":"inconnue","harnais":null,"configurations_mesurees":2,"date":"2025-07-03","source":"https://epoch.ai/benchmarks"},{"rang":27,"modele":"claude-opus-4","nom":"Claude Opus 4","editeur":"Anthropic","page":"https://quelleia.com/modeles/claude-opus-4/","score_publie":0.0416666666666666,"score_0_1":0.0694,"niveau_suggere":76.5,"reflexion":"budget","harnais":null,"configurations_mesurees":2,"date":"2025-07-01","source":"https://epoch.ai/benchmarks"},{"rang":27,"modele":"gemini-2-5-flash-jun-2025","nom":"Gemini 2.5 Flash (Jun 2025)","editeur":"Google DeepMind","page":"https://quelleia.com/modeles/gemini-2-5-flash-jun-2025/","score_publie":0.0416666666666666,"score_0_1":0.0694,"niveau_suggere":76.5,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2025-12-18","source":"https://epoch.ai/benchmarks"},{"rang":27,"modele":"o3-mini","nom":"o3-mini","editeur":"OpenAI","page":"https://quelleia.com/modeles/o3-mini/","score_publie":0.0416666666666666,"score_0_1":0.0694,"niveau_suggere":76.5,"reflexion":"elevee","harnais":null,"configurations_mesurees":1,"date":"2025-07-01","source":"https://epoch.ai/benchmarks"},{"rang":36,"modele":"glm-4-6","nom":"GLM-4.6","editeur":"Z.ai (Zhipu AI)","page":"https://quelleia.com/modeles/glm-4-6/","score_publie":0.0212765957446808,"score_0_1":0.0355,"niveau_suggere":72.8,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2025-12-08","source":"https://epoch.ai/benchmarks"},{"rang":37,"modele":"glm-5","nom":"GLM-5","editeur":"Z.ai (Zhipu AI)","page":"https://quelleia.com/modeles/glm-5/","score_publie":0.021,"score_0_1":0.035,"niveau_suggere":72.7,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2026-02-19","source":"https://epoch.ai/benchmarks"},{"rang":37,"modele":"deepseek-v3-2","nom":"DeepSeek-V3.2","editeur":"DeepSeek","page":"https://quelleia.com/modeles/deepseek-v3-2/","score_publie":0.021,"score_0_1":0.035,"niveau_suggere":72.7,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2025-12-16","source":"https://epoch.ai/benchmarks"},{"rang":39,"modele":"o3","nom":"o3","editeur":"OpenAI","page":"https://quelleia.com/modeles/o3/","score_publie":0.0208333333333333,"score_0_1":0.0347,"niveau_suggere":72.7,"reflexion":"elevee","harnais":null,"configurations_mesurees":1,"date":"2025-07-01","source":"https://epoch.ai/benchmarks"},{"rang":39,"modele":"grok-4","nom":"Grok 4","editeur":"xAI","page":"https://quelleia.com/modeles/grok-4/","score_publie":0.0208333333333333,"score_0_1":0.0347,"niveau_suggere":72.7,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2025-08-11","source":"https://epoch.ai/benchmarks"},{"rang":39,"modele":"qwen-3-5-plus-hosted-397b-a17b","nom":"Qwen 3.5 Plus (hosted 397B-A17B)","editeur":"Alibaba","page":"https://quelleia.com/modeles/qwen-3-5-plus-hosted-397b-a17b/","score_publie":0.0208333333333333,"score_0_1":0.0347,"niveau_suggere":72.7,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2026-05-15","source":"https://epoch.ai/benchmarks"},{"rang":39,"modele":"claude-haiku-4-5","nom":"Claude Haiku 4.5","editeur":"Anthropic","page":"https://quelleia.com/modeles/claude-haiku-4-5/","score_publie":0.0208333333333333,"score_0_1":0.0347,"niveau_suggere":72.7,"reflexion":"budget","harnais":null,"configurations_mesurees":1,"date":"2025-10-22","source":"https://epoch.ai/benchmarks"},{"rang":39,"modele":"gpt-5-nano","nom":"GPT-5 nano","editeur":"OpenAI","page":"https://quelleia.com/modeles/gpt-5-nano/","score_publie":0.0208333333333333,"score_0_1":0.0347,"niveau_suggere":72.7,"reflexion":"moyenne","harnais":null,"configurations_mesurees":2,"date":"2025-08-07","source":"https://epoch.ai/benchmarks"},{"rang":44,"modele":"gpt-5-4-mini","nom":"GPT-5.4 Mini","editeur":"OpenAI","page":"https://quelleia.com/modeles/gpt-5-4-mini/","score_publie":0.0208,"score_0_1":0.0347,"niveau_suggere":72.7,"reflexion":"elevee","harnais":null,"configurations_mesurees":1,"date":"2026-04-15","source":"https://epoch.ai/benchmarks"},{"rang":44,"modele":"grok-4-heavy","nom":"Grok 4 Heavy","editeur":"xAI","page":"https://quelleia.com/modeles/grok-4-heavy/","score_publie":0.0208,"score_0_1":0.0347,"niveau_suggere":72.7,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2025-10-09","source":"https://epoch.ai/benchmarks"},{"rang":46,"modele":"glm-4-7","nom":"GLM-4.7","editeur":"Z.ai (Zhipu AI)","page":"https://quelleia.com/modeles/glm-4-7/","score_publie":0,"score_0_1":0,"niveau_suggere":66,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2026-01-30","source":"https://epoch.ai/benchmarks"},{"rang":46,"modele":"kimi-k2-thinking","nom":"Kimi K2 Thinking","editeur":"Moonshot","page":"https://quelleia.com/modeles/kimi-k2-thinking/","score_publie":0,"score_0_1":0,"niveau_suggere":66,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2025-12-04","source":"https://epoch.ai/benchmarks"},{"rang":46,"modele":"qwen-3-6-flash","nom":"Qwen 3.6 Flash","editeur":"Alibaba","page":"https://quelleia.com/modeles/qwen-3-6-flash/","score_publie":0,"score_0_1":0,"niveau_suggere":66,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2026-05-12","source":"https://epoch.ai/benchmarks"},{"rang":46,"modele":"claude-sonnet-4","nom":"Claude Sonnet 4","editeur":"Anthropic","page":"https://quelleia.com/modeles/claude-sonnet-4/","score_publie":0,"score_0_1":0,"niveau_suggere":66,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2025-07-01","source":"https://epoch.ai/benchmarks"},{"rang":46,"modele":"qwen-3-5-flash-hosted-35b-a3b","nom":"Qwen 3.5 Flash (hosted 35B-A3B)","editeur":"Alibaba","page":"https://quelleia.com/modeles/qwen-3-5-flash-hosted-35b-a3b/","score_publie":0,"score_0_1":0,"niveau_suggere":66,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2026-05-12","source":"https://epoch.ai/benchmarks"},{"rang":46,"modele":"qwen3-235b-a22b-thinking-jul-2025","nom":"Qwen3-235B-A22B-Thinking (Jul 2025)","editeur":"Alibaba","page":"https://quelleia.com/modeles/qwen3-235b-a22b-thinking-jul-2025/","score_publie":0,"score_0_1":0,"niveau_suggere":66,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2025-12-11","source":"https://epoch.ai/benchmarks"},{"rang":46,"modele":"grok-3","nom":"Grok 3","editeur":"xAI","page":"https://quelleia.com/modeles/grok-3/","score_publie":0,"score_0_1":0,"niveau_suggere":66,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2025-07-01","source":"https://epoch.ai/benchmarks"},{"rang":46,"modele":"gpt-4-1","nom":"GPT-4.1","editeur":"OpenAI","page":"https://quelleia.com/modeles/gpt-4-1/","score_publie":0,"score_0_1":0,"niveau_suggere":66,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2025-07-01","source":"https://epoch.ai/benchmarks"},{"rang":46,"modele":"claude-3-5-sonnet","nom":"Claude 3.5 Sonnet","editeur":"Anthropic","page":"https://quelleia.com/modeles/claude-3-5-sonnet/","score_publie":0,"score_0_1":0,"niveau_suggere":66,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2025-07-01","source":"https://epoch.ai/benchmarks"},{"rang":46,"modele":"claude-3-5-sonnet-october-2024","nom":"Claude 3.5 Sonnet (October 2024)","editeur":"Anthropic","page":"https://quelleia.com/modeles/claude-3-5-sonnet-october-2024/","score_publie":0,"score_0_1":0,"niveau_suggere":66,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":"2025-07-01","source":"https://epoch.ai/benchmarks"}]}