China's MiniMax plans 2.7 trillion-parameter Pro model, possibly the largest open-weight model ever
koltregaskes · x · 2026-10-19
Per The Information and Reuters, MiniMax is preparing to release a Pro model with a total of 2.7 trillion parameters. Reuters (July 8, 2026), citing a person with direct knowledge, says it would be the largest open-weight model released by a Chinese firm and possibly the largest in the world.
- The trillion-parameter race is driven by demand for autonomous systems doing complex reasoning
- Analysts say the threshold is critical for multi-step systems operating without human intervention
More from Models
- inclusionAI Releases Ming-3.0-audio-flash: 126B MoE Audio Model, MIT — BerryBlush_Rogers691 · 2026-10-19
- inclusionAI open-sources Ming-3.0-audio-tiny: 9.3B MoE audio-language model (MIT) — jacek2023 · 2026-10-19
- 'Pre-training is definitely not dead,' dev says amid debate over GLM's frontier ambitions — himanshustwts · 2026-10-19
- Meta's Muse Spark Open: 816B Open-Weights Model Tested Against Beam — Sam Witteveen · 2026-10-19
- Anthropic, Google and OpenAI are all preparing new AI models, The Information reports — coinfanking · 2026-10-19
- Nace AI launches Drex Max, a screenshot-in decision model for agents at 174 ms and $0.06/M tokens — testingcatalog · 2026-10-19