{"source":"Quelle IA, le classement des modèles de langage","page":"https://quelleia.com/benchmarks/lambada/","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":"lambada","nom":"LAMBADA","tache":"autre","etat":"archive","echelle":"unite","modeles":26,"poids_general":0,"difficulte":-111.4,"pente":0.0096,"en_une_phrase":"Deviner le dernier mot d'un passage de roman, ce qui exige d'avoir suivi tout le paragraphe.","notation":"Part de mots exactement retrouvés.","limites":"Test de 2016, pensé pour les modèles d'avant ChatGPT. Les scores viennent d'articles anciens aux réglages variés.","mainteneur":"Paperno et al. (université de Trente, 2016)","url":"https://arxiv.org/abs/1606.06031","verifie":true},"tache":{"id":"autre","nom":"Tests historiques","poids_general":0,"benchmarks_actifs":0},"modeles_mesures":26,"mesures":53,"classement":[{"rang":1,"modele":"megatron-turing-nlg-530b","nom":"Megatron-Turing NLG 530B","editeur":"Microsoft","page":"https://quelleia.com/modeles/megatron-turing-nlg-530b/","score_publie":0.8715,"score_0_1":0.8715,"niveau_suggere":88.3,"reflexion":"inconnue","harnais":null,"configurations_mesurees":4,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":2,"modele":"instructgpt-175b","nom":"InstructGPT 175B","editeur":"OpenAI","page":"https://quelleia.com/modeles/instructgpt-175b/","score_publie":0.864,"score_0_1":0.864,"niveau_suggere":81.5,"reflexion":"inconnue","harnais":null,"configurations_mesurees":6,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":3,"modele":"falcon-180b","nom":"Falcon-180B","editeur":"Technology Innovation Institute","page":"https://quelleia.com/modeles/falcon-180b/","score_publie":0.798,"score_0_1":0.798,"niveau_suggere":31.9,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":4,"modele":"llama-2-70b","nom":"Llama 2-70B","editeur":"Meta AI","page":"https://quelleia.com/modeles/llama-2-70b/","score_publie":0.789,"score_0_1":0.789,"niveau_suggere":26.2,"reflexion":"inconnue","harnais":null,"configurations_mesurees":2,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":5,"modele":"inflection-1","nom":"Inflection-1","editeur":"Inflection AI","page":"https://quelleia.com/modeles/inflection-1/","score_publie":0.785,"score_0_1":0.785,"niveau_suggere":23.7,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":6,"modele":"palm-540b","nom":"PaLM (540B)","editeur":"Google Research","page":"https://quelleia.com/modeles/palm-540b/","score_publie":0.779,"score_0_1":0.779,"niveau_suggere":20,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":7,"modele":"llama-65b","nom":"LLaMA-65B","editeur":"Meta AI","page":"https://quelleia.com/modeles/llama-65b/","score_publie":0.777,"score_0_1":0.777,"niveau_suggere":18.8,"reflexion":"inconnue","harnais":null,"configurations_mesurees":2,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":8,"modele":"chinchilla","nom":"Chinchilla","editeur":"DeepMind","page":"https://quelleia.com/modeles/chinchilla/","score_publie":0.774,"score_0_1":0.774,"niveau_suggere":17,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":9,"modele":"falcon-40b","nom":"Falcon-40B","editeur":"Technology Innovation Institute","page":"https://quelleia.com/modeles/falcon-40b/","score_publie":0.773,"score_0_1":0.773,"niveau_suggere":16.4,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":10,"modele":"llama-33b","nom":"LLaMA-33B","editeur":"Meta AI","page":"https://quelleia.com/modeles/llama-33b/","score_publie":0.772,"score_0_1":0.772,"niveau_suggere":15.8,"reflexion":"inconnue","harnais":null,"configurations_mesurees":2,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":11,"modele":"llama-2-13b","nom":"Llama 2-13B","editeur":"Meta AI","page":"https://quelleia.com/modeles/llama-2-13b/","score_publie":0.765,"score_0_1":0.765,"niveau_suggere":11.7,"reflexion":"inconnue","harnais":null,"configurations_mesurees":2,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":12,"modele":"llama-13b","nom":"LLaMA-13B","editeur":"Meta AI","page":"https://quelleia.com/modeles/llama-13b/","score_publie":0.752,"score_0_1":0.752,"niveau_suggere":4.3,"reflexion":"inconnue","harnais":null,"configurations_mesurees":2,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":13,"modele":"falcon-7b","nom":"Falcon-7B","editeur":"Technology Innovation Institute","page":"https://quelleia.com/modeles/falcon-7b/","score_publie":0.749,"score_0_1":0.749,"niveau_suggere":2.7,"reflexion":"inconnue","harnais":null,"configurations_mesurees":1,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":14,"modele":"gopher-280b","nom":"Gopher (280B)","editeur":"DeepMind","page":"https://quelleia.com/modeles/gopher-280b/","score_publie":0.745,"score_0_1":0.745,"niveau_suggere":0.4,"reflexion":"inconnue","harnais":null,"configurations_mesurees":3,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":15,"modele":"baichuan2-13b","nom":"Baichuan2-13B","editeur":"Baichuan","page":"https://quelleia.com/modeles/baichuan2-13b/","score_publie":0.74,"score_0_1":0.74,"niveau_suggere":-2.3,"reflexion":"inconnue","harnais":null,"configurations_mesurees":2,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":16,"modele":"llama-2-7b","nom":"Llama 2-7B","editeur":"Meta AI","page":"https://quelleia.com/modeles/llama-2-7b/","score_publie":0.733,"score_0_1":0.733,"niveau_suggere":-6,"reflexion":"inconnue","harnais":null,"configurations_mesurees":2,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":16,"modele":"baichuan-2-7b","nom":"Baichuan 2-7B","editeur":"Baichuan","page":"https://quelleia.com/modeles/baichuan-2-7b/","score_publie":0.733,"score_0_1":0.733,"niveau_suggere":-6,"reflexion":"inconnue","harnais":null,"configurations_mesurees":2,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":16,"modele":"llama-7b","nom":"LLaMA-7B","editeur":"Meta AI","page":"https://quelleia.com/modeles/llama-7b/","score_publie":0.733,"score_0_1":0.733,"niveau_suggere":-6,"reflexion":"inconnue","harnais":null,"configurations_mesurees":2,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":19,"modele":"internlm-20b","nom":"internlm-20b","editeur":"Shanghai AI Laboratory","page":"https://quelleia.com/modeles/internlm-20b/","score_publie":0.718,"score_0_1":0.718,"niveau_suggere":-13.9,"reflexion":"inconnue","harnais":null,"configurations_mesurees":2,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":20,"modele":"stable-beluga-2","nom":"Stable Beluga 2","editeur":"Stability AI","page":"https://quelleia.com/modeles/stable-beluga-2/","score_publie":0.713,"score_0_1":0.713,"niveau_suggere":-16.5,"reflexion":"inconnue","harnais":null,"configurations_mesurees":2,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":21,"modele":"qwen-14b","nom":"Qwen-14B","editeur":"Alibaba","page":"https://quelleia.com/modeles/qwen-14b/","score_publie":0.711,"score_0_1":0.711,"niveau_suggere":-17.5,"reflexion":"inconnue","harnais":null,"configurations_mesurees":2,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":22,"modele":"mpt-7b","nom":"MPT-7B","editeur":"MosaicML","page":"https://quelleia.com/modeles/mpt-7b/","score_publie":0.7,"score_0_1":0.7,"niveau_suggere":-23,"reflexion":"inconnue","harnais":null,"configurations_mesurees":2,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":23,"modele":"qwen-7b","nom":"Qwen-7B","editeur":"Alibaba","page":"https://quelleia.com/modeles/qwen-7b/","score_publie":0.679,"score_0_1":0.679,"niveau_suggere":-33.2,"reflexion":"inconnue","harnais":null,"configurations_mesurees":2,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":24,"modele":"internlm-7b","nom":"internlm-7b","editeur":"Shanghai AI Laboratory","page":"https://quelleia.com/modeles/internlm-7b/","score_publie":0.67,"score_0_1":0.67,"niveau_suggere":-37.5,"reflexion":"inconnue","harnais":null,"configurations_mesurees":2,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":25,"modele":"qwen-1-8b","nom":"Qwen-1_8B","editeur":"Alibaba","page":"https://quelleia.com/modeles/qwen-1-8b/","score_publie":0.584,"score_0_1":0.584,"niveau_suggere":-76,"reflexion":"inconnue","harnais":null,"configurations_mesurees":2,"date":null,"source":"https://epoch.ai/benchmarks"},{"rang":26,"modele":"chatglm2-6b","nom":"chatglm2-6b","editeur":"Z.ai (Zhipu AI)","page":"https://quelleia.com/modeles/chatglm2-6b/","score_publie":0.543,"score_0_1":0.543,"niveau_suggere":-93.4,"reflexion":"inconnue","harnais":null,"configurations_mesurees":2,"date":null,"source":"https://epoch.ai/benchmarks"}]}