{"source":"Quelle IA, le classement des modèles de langage","page":"https://quelleia.com/benchmarks/the_agent_company/","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":"the_agent_company","nom":"TheAgentCompany","tache":"agents","etat":"archive","echelle":"unite","modeles":14,"poids_general":0,"difficulte":76.3,"pente":0.1024,"en_une_phrase":"175 tâches de bureau dans une fausse entreprise de logiciel : coder, gérer un projet, traiter des dossiers de RH ou de finance.","notation":"Part de tâches entièrement accomplies ; un second score accorde des points partiels par étape franchie.","limites":"Peu de modèles testés, et aucun sorti après septembre 2025 dans les données d'Epoch.","mainteneur":"Université Carnegie Mellon","url":"https://the-agent-company.com/","verifie":true},"tache":{"id":"agents","nom":"Agents","poids_general":15,"benchmarks_actifs":9},"modeles_mesures":14,"mesures":16,"classement":[{"rang":1,"modele":"deepseek-v3-2-exp","nom":"DeepSeek-V3.2-Exp","editeur":"DeepSeek","page":"https://quelleia.com/modeles/deepseek-v3-2-exp/","score_publie":0.429,"score_0_1":0.429,"niveau_suggere":73.5,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":2,"modele":"gemini-2-5-flash-sep-2025","nom":"Gemini 2.5 Flash (Sep 2025)","editeur":"Google DeepMind","page":"https://quelleia.com/modeles/gemini-2-5-flash-sep-2025/","score_publie":0.411,"score_0_1":0.411,"niveau_suggere":72.8,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":3,"modele":"claude-sonnet-4","nom":"Claude Sonnet 4","editeur":"Anthropic","page":"https://quelleia.com/modeles/claude-sonnet-4/","score_publie":0.331,"score_0_1":0.331,"niveau_suggere":69.4,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":4,"modele":"claude-3-7-sonnet","nom":"Claude 3.7 Sonnet","editeur":"Anthropic","page":"https://quelleia.com/modeles/claude-3-7-sonnet/","score_publie":0.309,"score_0_1":0.309,"niveau_suggere":68.4,"reflexion":"inconnue","harnais":null,"configurations_mesurees":2,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":5,"modele":"gemini-2-5-pro-may-2025","nom":"Gemini 2.5 Pro (May 2025)","editeur":"Google DeepMind","page":"https://quelleia.com/modeles/gemini-2-5-pro-may-2025/","score_publie":0.303,"score_0_1":0.303,"niveau_suggere":68.2,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":6,"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":0.24,"score_0_1":0.24,"niveau_suggere":65,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":7,"modele":"gemini-2-0-flash-feb-2025","nom":"Gemini 2.0 Flash (Feb 2025)","editeur":"Google DeepMind","page":"https://quelleia.com/modeles/gemini-2-0-flash-feb-2025/","score_publie":0.114,"score_0_1":0.114,"niveau_suggere":56.3,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":8,"modele":"gpt-4o-nov-2024","nom":"GPT-4o (Nov 2024)","editeur":"OpenAI","page":"https://quelleia.com/modeles/gpt-4o-nov-2024/","score_publie":0.086,"score_0_1":0.086,"niveau_suggere":53.2,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":9,"modele":"llama-3-1-405b","nom":"Llama 3.1-405B","editeur":"Meta AI","page":"https://quelleia.com/modeles/llama-3-1-405b/","score_publie":0.074,"score_0_1":0.074,"niveau_suggere":51.6,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":10,"modele":"llama-3-1-70b","nom":"Llama 3.1-70B","editeur":"Meta AI","page":"https://quelleia.com/modeles/llama-3-1-70b/","score_publie":0.069,"score_0_1":0.069,"niveau_suggere":50.9,"reflexion":"inconnue","harnais":null,"configurations_mesurees":2,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":11,"modele":"qwen2-5-72b","nom":"Qwen2.5-72B","editeur":"Alibaba","page":"https://quelleia.com/modeles/qwen2-5-72b/","score_publie":0.057,"score_0_1":0.057,"niveau_suggere":48.9,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":12,"modele":"gemini-1-5-pro-sept-2024","nom":"Gemini 1.5 Pro (Sept 2024)","editeur":"Google DeepMind","page":"https://quelleia.com/modeles/gemini-1-5-pro-sept-2024/","score_publie":0.034,"score_0_1":0.034,"niveau_suggere":43.6,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":13,"modele":"amazon-nova-pro","nom":"Amazon Nova Pro","editeur":"Amazon","page":"https://quelleia.com/modeles/amazon-nova-pro/","score_publie":0.017,"score_0_1":0.017,"niveau_suggere":36.7,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":14,"modele":"qwen2-72b","nom":"Qwen2-72B","editeur":"Alibaba","page":"https://quelleia.com/modeles/qwen2-72b/","score_publie":0.011,"score_0_1":0.011,"niveau_suggere":32.4,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":null,"source":"https://epoch.ai/benchmarks"}]}