Mastodawn

36Kr Japan | 最大級の中国テック・スタートアップ専門メディア Mar 13

中国の生成AI「MiniMax」、上場後初の決算　売上159％増と巨額赤字の裏側

https://fed.brid.gy/r/https://36kr.jp/461747/

36Kr Japan | 最大級の中国テック・スタートアップ専門メディア Mar 4

中国製AIモデル、API利用量で初めて米国を逆転　トップ5の4枠握る

https://fed.brid.gy/r/https://36kr.jp/460418/

Ellie‽ (Radio Edit)Feb 22

What's The Deal With "THAT" CONCRETE Section of the M25 MOTORWAY?

Oh god, I hate driving on that bit. Well, I hate driving on all bits of the M25 to be fair, but that bit in particular is 'orrible.

https://youtu.be/_YBY8EHkZK8

#M25 #Motorways #UKRoads #UKMotorways #AutoShenanigans

What's The Deal With "THAT" CONCRETE Section of the M25 MOTORWAY?

YouTube

AI Sparkup Feb 14

MiniMax M2.5, 시간당 1달러로 실행하는 코딩 에이전트

MiniMax M2.5는 시간당 1달러로 실행 가능한 코딩 에이전트입니다. SWE-Bench 80.2% 달성하며 실무 도입의 경제적 장벽을 낮춥니다.

https://aisparkup.com/posts/9245

GOMOOT

Feb 13

📌 MiniMax M2.5: il modello open che sfida Claude Opus 4.6 a un ventesimo del prezzo

https://gomoot.com/minimax-m2-5-il-modello-open-che-sfida-claude-opus-4-6-a-un-ventesimo-del-prezzo/

#ai #m25 #minimax #news #tech @MiniMax_AI

eicker.news ᳇ tech news Feb 13

#MiniMax releases #M25 AI model, claiming "intelligence too cheap to meter" at just $0.30/1M input and $1.20/1M output tokens. The model achieves #SOTA results in #coding (80.2% SWE-Bench Verified), #agentic tool use, and #officework, running 37% faster than M2.1. https://www.minimax.io/news/minimax-m25?eicker.news #tech #media #news

MiniMax M2.5: 更快更强更智能，为真实世界生产力而生

MiniMax

sayzard Feb 13

cedric (@cedric_chee)

MiniMax M2.5는 총 230B 파라미터에 활성 파라미터가 단 10B로 작아 홈랩에서 유력한 후보라는 평가입니다. M2 시리즈는 LocalLLaMA 커뮤니티에서 인기를 끌고 있으며, 작성자는 추론 제공자들이 TPS를 극한까지 끌어올려 성능을 보여주길 바란다고 언급했습니다.

https://x.com/cedric_chee/status/2022148184508969164

#minimax #m25 #homelab #localllama

cedric (@cedric_chee) on X

MiniMax M2.5 is surprisingly small for what it delivers: 230B total, just 10B active. That makes it a serious candidate for home labs. M2 series has been big with the LocalLLaMA crowd. Would love to see inference providers push its TPS to the limit. ⚡️⚡️⚡️

X (formerly Twitter)

AIagent.at 🤖 AI News Feb 12

#MiniMax releases #M25 AI model, claiming "intelligence too cheap to meter" at just $0.30/1M input and $1.20/1M output tokens. The model achieves SOTA results in coding (80.2% SWE-Bench Verified), agentic tool use, and office work, running 37% faster than M2.1. https://www.minimax.io/news/minimax-m25?AIagents.at #AIagent #AI #ML #NLP #LLM #GenAI