MiniMax has launched its M3 model, featuring a 1 million token context window and a Sparse Attention architecture. This design significantly accelerates response generation, reportedly by over 15 times. The M3 model is notable for being an open-weight model that effectively combines multimodal capabilities with strong engineering features. AI
IMPACT Sets new SOTA on context window length and multimodal engineering.
RANK_REASON Frontier-lab model release with system card. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
Read on Mastodon — fosstodon.org →
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →