Video-MME
Resultados de benchmarks de IA e desempenho dos modelos
Tipo: percentageMultimodal
Resultados registrados
29
Pontuação média
77.03%
Faixa de pontuação
55% - 89.50%
Modelo líder
89.50% - Seed 2.0 Pro
Pontuações ao longo do tempo
Pontuações individuais do benchmark exibidas por data.
Resultados de modelos
| Organização | Modelos | Data informada | Maior pontuação | Informações | Declarado pelo provedor | Fonte |
|---|---|---|---|---|---|---|
| Seed 2.0 Pro | 14 de fev. de 2026 | 89,50% | Seed2 official benchmark table | VideoMME ‡ | Sim | Fonte | |
| Seed 2.0 Lite | 14 de fev. de 2026 | 87,70% | Seed2 official benchmark table | VideoMME ‡ | Sim | Fonte | |
| MiMo V2.5 | 23 de abr. de 2026 | 87,70% | Video understanding benchmark from official MiMo-V2.5 model card image | Sim | Fonte | |
| Kimi K2.5 | 27 de jan. de 2026 | 87,40% | - | Sim | Fonte | |
| Gemini 2.5 Computer Use Preview | 07 de out. de 2025 | 84,80% | inferred family alias from gemini-2.5-pro (score=0.3960; benches=16) | Sim | Fonte | |
| Gemini 2.5 Pro Experimental (2025-03-25) | 25 de mar. de 2025 | 84,80% | inferred alias from gemini-2.5-pro | Sim | Fonte | |
| Gemini 2.5 Pro Preview TTS (2025-12-10) | 10 de dez. de 2025 | 84,80% | inferred modality/version alias from gemini-2.5-pro | Sim | Fonte | |
| Gemini Embedding 2 Preview | 10 de mar. de 2026 | 84,80% | manual fallback alias from gemini-2.5-pro | Sim | Fonte | |
| Qwen 3.6 Plus | 01 de abr. de 2026 | 84,20% | - | Sim | Fonte | |
| Seed 2.0 Mini | 14 de fev. de 2026 | 81,20% | Seed2 official benchmark table | VideoMME ‡ | Sim | Fonte | |
| Gemini 1.5 Pro 001 | 23 de mai. de 2024 | 78,60% | inferred alias from gemini-1.5-pro | Sim | Fonte | |
| Gemini 1.5 Pro Exp (2024-08-01) | 01 de ago. de 2024 | 78,60% | inferred alias from gemini-1.5-pro | Sim | Fonte | |
| Gemini 1.5 Pro Exp (2024-08-27) | 27 de ago. de 2024 | 78,60% | inferred alias from gemini-1.5-pro | Sim | Fonte | |
| Gemini Robotics ER 1.5 Preview | 25 de set. de 2025 | 78,60% | inferred family alias from gemini-1.5-pro (score=0.3717; benches=23) | Sim | Fonte | |
| LearnLM 1.5 Pro Experimental | 19 de nov. de 2024 | 78,60% | inferred family alias from gemini-1.5-pro (score=0.3700; benches=23) | Sim | Fonte | |
| Gemini 1.5 Flash 001 | 23 de mai. de 2024 | 76,10% | inferred alias from gemini-1.5-flash | Sim | Fonte | |
| Gemini 1.5 Flash Preview | 14 de mai. de 2024 | 76,10% | inferred alias from gemini-1.5-flash | Sim | Fonte | |
| Qwen 3 VL 30B A3B Instruct | - | 74,50% | - | Sim | - | |
| Qwen 3 VL 30B A3B Thinking | - | 73,30% | - | Sim | - | |
| Qwen 3 8B | - | 71,80% | inferred high-confidence family alias from qwen3-vl-8b-thinking (score=0.4600; benches=50) | Sim | - | |
| Qwen 3 Guard Stream 8B | - | 71,80% | inferred family alias from qwen3-vl-8b-thinking (score=0.3371; benches=50) | Sim | - | |
| Qwen 3 VL 8B Thinking | - | 71,80% | - | Sim | - | |
| Qwen 3 VL Embedding 8B | - | 71,80% | inferred high-confidence family alias from qwen3-vl-8b-thinking (score=0.5232; benches=50) | Sim | - | |
| Qwen 3 VL Reranker 8B | - | 71,80% | inferred high-confidence family alias from qwen3-vl-8b-thinking (score=0.5275; benches=50) | Sim | - | |
| Qwen 3 VL 8B Instruct | - | 71,40% | - | Sim | - |
