DocVQA
Résultats des évaluations de l’IA et performances des modèles
Type: percentage
Résultats enregistrés
44
Score moyen
93.41%
Plage des scores
85.60% - 96.40%
Modèle en tête
96.40% - Qwen 2.5 Math 72B
Scores au fil du temps
Scores individuels du benchmark tracés par date.
Résultats des modèles
| Organisation | Modèles | Date du résultat | Meilleur score | Informations | Déclaré par le fournisseur | Source |
|---|---|---|---|---|---|---|
| Qwen 2.5 Math 72B | - | 96,40 % | inferred high-confidence family alias from qwen2.5-vl-72b (score=0.4700; benches=30) | Oui | Source | |
| Qwen 2.5 Math PRM 72B | - | 96,40 % | inferred family alias from qwen2.5-vl-72b (score=0.4050; benches=30) | Oui | Source | |
| Qwen 2.5 Math RM 72B | - | 96,40 % | inferred family alias from qwen2.5-vl-72b (score=0.4092; benches=30) | Oui | Source | |
| Qwen 2.5 VL 72B Instruct | 26 janv. 2025 | 96,40 % | - | Oui | Source | |
| Qwen 2.5 0.5B | - | 95,70 % | inferred family alias from qwen2.5-vl-7b (score=0.3131; benches=32) | Oui | Source | |
| Qwen 2.5 1.5B | - | 95,70 % | inferred family alias from qwen2.5-vl-7b (score=0.3131; benches=32) | Oui | Source | |
| Qwen 2.5 3B | - | 95,70 % | inferred family alias from qwen2.5-vl-7b (score=0.3531; benches=32) | Oui | Source | |
| Qwen 2.5 VL 7B Instruct | - | 95,70 % | - | Oui | Source | |
| DeepSeek V4.1 Flash | 10 sept. 2026 | 95,60 % | Base model; 4-shot; LLM-Judge | Oui | Source | |
| Qwen 2.5 Coder 3B | - | 95,20 % | inferred family alias from qwen2.5-omni-7b (score=0.3000; benches=45) | Oui | Source | |
| Qwen 2.5 Coder 7B | - | 95,20 % | inferred high-confidence family alias from qwen2.5-omni-7b (score=0.4700; benches=45) | Oui | Source | |
| Qwen 2.5 Math 7B | - | 95,20 % | inferred high-confidence family alias from qwen2.5-omni-7b (score=0.4767; benches=45) | Oui | Source | |
| Qwen 2.5 Math 7B PRM800K | - | 95,20 % | inferred family alias from qwen2.5-omni-7b (score=0.3696; benches=45) | Oui | Source | |
| Qwen 2.5 Math PRM 7B | - | 95,20 % | inferred family alias from qwen2.5-omni-7b (score=0.4092; benches=45) | Oui | Source | |
| Qwen 2.5 Omni 3B | - | 95,20 % | inferred high-confidence family alias from qwen2.5-omni-7b (score=0.4933; benches=45) | Oui | Source | |
| Qwen 2.5 Omni 7B | - | 95,20 % | - | Oui | Source | |
| Mistral Small 3.2 | 20 juin 2025 | 94,86 % | - | Oui | Source | |
| Qwen 2.5 Coder 32B Instruct | - | 94,80 % | inferred high-confidence family alias from qwen2.5-vl-32b (score=0.4641; benches=28) | Oui | Source | |
| Qwen 2.5 VL 32B Instruct | 24 mars 2025 | 94,80 % | - | Oui | Source | |
| Qwen 2.5 VL 3B Instruct | - | 94,80 % | inferred high-confidence family alias from qwen2.5-vl-32b (score=0.4914; benches=28) | Oui | Source | |
| Llama 4 Maverick | 05 avr. 2025 | 94,40 % | - | Oui | Source | |
| Llama 4 Scout | 05 avr. 2025 | 94,40 % | - | Oui | Source | |
| Grok 2 | 13 août 2024 | 93,60 % | - | Oui | Source | |
| DeepSeek VL2 | 13 déc. 2024 | 93,30 % | - | Oui | Source | |
| Pixtral Large | 18 nov. 2024 | 93,30 % | - | Oui | Source |