{"version":1,"type":"story","url":"https://digestai.news/story/minimax-releases-m3-1-flash-model-at-0-15-per-1m-input-tokens","json":"https://digestai.news/story/minimax-releases-m3-1-flash-model-at-0-15-per-1m-input-tokens.json","markdown":"https://digestai.news/story/minimax-releases-m3-1-flash-model-at-0-15-per-1m-input-tokens.md","slug":"minimax-releases-m3-1-flash-model-at-0-15-per-1m-input-tokens","headline":"MiniMax releases M3.1 Flash model at $0.15 per 1M input tokens","summary":"MiniMax launched its M3.1 Flash model on October 19, 2026, making it available via its API, M Plan, and MiniMax Code desktop app. The model has 428 billion total parameters with about 23 billion active parameters in a mixture-of-experts architecture and supports a 1 million token context window. Pricing is set at $0.15 per million input tokens and $0.60 per million output tokens, with cache reads at $0.015 per million tokens. Model weights are scheduled for release on October 23, 2026, though the license has not been disclosed. MiniMax reports the model achieves a score of 62.2 on Automation-Bench v1.0.6, which it states is the top score in its internal benchmark table, while noting all performance figures are self-reported and lack independent verification.","keyPoints":["MiniMax M3.1 Flash has 428B total parameters, ~23B active, and 1M token context window","Pricing is $0.15 input and $0.60 output per 1M tokens, half the rate of M3 model","Model weights to be released October 23, 2026; reasoning cannot be disabled"],"whyItMatters":"The model offers near-frontier agentic performance at a low cost, potentially expanding access to powerful AI for developers and businesses using MiniMax’s platform.","category":{"slug":"models","name":"Generative AI & Models","url":"https://digestai.news/category/models"},"entities":{"companies":["MiniMax"],"models":["MiniMax M3.1 Flash","M3","Opus 5","GPT-5.6 Sol","DeepSeek-V4.1-Flash-1017","Kimi K3"],"people":["Asif Razzaq"]},"firstPublishedAt":"2026-10-19T12:37:58Z","updatedAt":"2026-10-19T13:26:44Z","sourceCount":2,"hasPrimarySource":false,"sources":[{"outlet":"MarkTechPost","title":"MiniMax Just Launched M3.1 Flash: 428B MoE Agentic Model at $0.15/M Input","url":"https://marktechpost.com/2026/10/19/minimax-m3-1-flash-release","publishedAt":"2026-10-19T12:37:58Z","type":"press","primary":false,"lead":true},{"outlet":"maccus.fr","title":"MiniMax launches M3.1 Flash at $0.15 per million input tokens, half the list price of M3","url":"https://maccus.fr/en/minimax-m3-1-flash-launch","publishedAt":"2026-10-19T13:26:44Z","type":"press","primary":false,"lead":false}],"sourceNotes":null,"discussions":[{"site":"Reddit","url":"https://www.reddit.com/r/LocalLLaMA/comments/1x9wv1x/minimax_launches_m31_flash_at_015_per_million/","points":null}],"thread":{"title":"AI Model Race Accelerates with New Releases","url":"https://digestai.news/thread/typesafe-ai-releases-jev-1-14-with-images-box-type-and-262k-token-context","storyCount":2},"cite":{"text":"Digest AI, \"MiniMax releases M3.1 Flash model at $0.15 per 1M input tokens\", 19 October 2026, https://digestai.news/story/minimax-releases-m3-1-flash-model-at-0-15-per-1m-input-tokens","publisher":"Digest AI","title":"MiniMax releases M3.1 Flash model at $0.15 per 1M input tokens","datePublished":"2026-10-19T12:37:58Z","url":"https://digestai.news/story/minimax-releases-m3-1-flash-model-at-0-15-per-1m-input-tokens"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}