An open-weight model just out-coded the giants. MiniMax M3 scored 59% on SWE-Bench Pro — beating GPT-5.5 and Gemini 3.1 Pro on real software engineering tasks. The part builders should care about: it runs a 1M-token context at ~1/20th the per-token compute of the last gen, with 9x faster prefill and 15x faster decode. Translation: huge-context AI just got way cheaper — and the weights are open. Will open models eat the closed frontier's lunch? www.contentbuffer.com #AI #OpenSource #Coding