« MiniMax-M1 » : différence entre les versions
(Page créée avec « ==en construction== == Définition == == Français == ''' MiniMax-M1''' == Anglais == '''MiniMax-M1''' MiniMax-M1, the world's first open-weight, large-scale hybrid-attention reasoning model. MiniMax-M1 is powered by a hybrid Mixture-of-Experts (MoE) architecture combined with a lightning attention mechanism. The model is developed based on our previous MiniMax-Text-01 model, which contains a total of 456 billion parameters with 45.9 billion parameters ac... ») |
Aucun résumé des modifications |
||
| Ligne 13 : | Ligne 13 : | ||
MiniMax-M1, the world's first open-weight, large-scale hybrid-attention reasoning model. MiniMax-M1 is powered by a hybrid Mixture-of-Experts (MoE) architecture combined with a lightning attention mechanism. The model is developed based on our previous MiniMax-Text-01 model, which contains a total of 456 billion parameters with 45.9 billion parameters activated per token. Consistent with MiniMax-Text-01, the M1 model natively supports a context length of 1 million tokens, 8x the context size of DeepSeek R1. Furthermore, the lightning attention mechanism in MiniMax-M1 enables efficient scaling of test-time compute – | MiniMax-M1, the world's first open-weight, large-scale hybrid-attention reasoning model. MiniMax-M1 is powered by a hybrid Mixture-of-Experts (MoE) architecture combined with a lightning attention mechanism. The model is developed based on our previous MiniMax-Text-01 model, which contains a total of 456 billion parameters with 45.9 billion parameters activated per token. Consistent with MiniMax-Text-01, the M1 model natively supports a context length of 1 million tokens, 8x the context size of DeepSeek R1. Furthermore, the lightning attention mechanism in MiniMax-M1 enables efficient scaling of test-time compute – | ||
== Sources == | |||
[https://arxiv.org/abs/2506.13585 Source : arxiv] | |||
[https://huggingface.co/MiniMaxAI/MiniMax-M1-80k Source : huggingface] | [https://huggingface.co/MiniMaxAI/MiniMax-M1-80k Source : huggingface] | ||
[https://minimax-m1.com/ Source : MiniMax-M1] | |||
[[Catégorie:vocabulary]] | [[Catégorie:vocabulary]] | ||
Version du 8 juillet 2025 à 13:17
en construction
Définition
Français
MiniMax-M1
Anglais
MiniMax-M1
MiniMax-M1, the world's first open-weight, large-scale hybrid-attention reasoning model. MiniMax-M1 is powered by a hybrid Mixture-of-Experts (MoE) architecture combined with a lightning attention mechanism. The model is developed based on our previous MiniMax-Text-01 model, which contains a total of 456 billion parameters with 45.9 billion parameters activated per token. Consistent with MiniMax-Text-01, the M1 model natively supports a context length of 1 million tokens, 8x the context size of DeepSeek R1. Furthermore, the lightning attention mechanism in MiniMax-M1 enables efficient scaling of test-time compute –
Sources
