MiniMax operates as a Chinese AI developer releasing open-weight generative models. Its most recent release is MiniMax M3.1 Flash, announced October 19, 2026, a model for agent and coding tasks with published benchmark scores and per-token pricing.
MiniMax's run of releases over the past months falls into two lines: video and music generation on one side, agent and coding models on the other. The video line starts with MiniMax H3, announced on August 9, 2026, with up to 15 seconds of audio-synced video from text, images, or voice, support for 768p or 2K resolution, and multiple simultaneous media inputs. In a blind test by Artificial Analysis, H3 ranked second for text-to-video, behind Gemini Omni Flash and ahead of Seedance 2.0. The model was available via API then, with open weights planned. The company released H3 as an open-weight model on August 3, 2026, and its Omni Reference feature accepts up to nine images, three videos, and three audio tracks, with local runs on an RTX 4090.
Community and cloud work followed. The AI cloud service fal built MiniMax H3 Max, which generates 15-second videos in about 13 seconds and offers 5-second videos free up to five times daily with Japanese prompt support, at $0.08 per second at 768p. H3 Max weights had not been released 28 days after fal said on X that it would publish them, and a first-person GameBusiness.jp report timed an RTX 5090 running H3 locally at 31.6 seconds for a 5-second image-to-video clip, down from 193.5 seconds in 41 days.
Separately, MiniMax released MiniMax-Music3 on August 14, 2026, a free music generation model that produces songs with vocals from text prompts, supports Japanese vocals, runs locally, and makes tracks up to five minutes long. The most recent release moves to agents. MiniMax M3.1 Flash, announced October 19, 2026, reports 61.7 on OSWorld V2, 29.6 on Agents' Last Exam, 76.2 on BankerToolBench, 73.7 percent on DeepSWE v1.1, and 82 percent on CyberGym, where MiniMax claims first among Chinese-made AI models. Output speed exceeds 200 tokens per second, and pricing is $0.15 per million input tokens and $0.60 per million output tokens.
Synthesized by Yomimono from the cited Yomimono stories below, each itself
sourced, then editorially reviewed. Every
fact links the story it came from.
4d ago
MiniMax announced MiniMax M3.1 Flash on October 19, 2026. It reports the model scores 76.2 on BankerToolBench, 73.7 percent on DeepSWE v1.1, 61.7 on OSWorld V2, 29.6 on Agents' Last Exam, and 82 percent on CyberGym, where MiniMax claims the model ranked first among Chinese-made AI models. Output speed exceeds 200 tokens per second, inference runs about three times faster than the previous model, and agent task costs fall 70 percent. Pricing is $0.15 per million input tokens and $0.60 per million output tokens, with cached input at $0.015 per million.
Oct 5
A first-person report on GameBusiness.jp timed an RTX 5090 running MiniMax H3 locally. One 5-second i2v clip went from 193.5 seconds to 31.6 seconds in 41 days, a 6.12x gain, using a 20-step baseline, the official Turbo 8-step LoRA, Hao AI Lab's FastH3 4-step, a community-converted TaoMate 3-step LoRA, INT8 attention, and sparse attention. H3 Max weights have not been released 28 days after fal said on X that it would publish them. Cloud H3 Max Turbo ran at 1.49 seconds on different hardware.
Aug 30
The video generation AI MiniMax H3, released as an open-weight model on August 3, has spawned rapid community updates. The AI cloud service fal developed MiniMax H3 Max, which generates 15-second videos in about 13 seconds while maintaining quality. It offers 5-second videos free up to five times daily, with Japanese prompt support, and costs $0.08 per second at 768p, a 50% off sale until September 1.
Aug 15
Chinese AI developer MiniMax released the music generation AI MiniMax-Music3 on August 14, 2026. The model generates songs with vocals from text prompts, supports Japanese vocals, and can produce tracks up to five minutes long. It is free to use and can run locally, with a demo app available on Hugging Face.
Aug 9
Chinese AI firm MiniMax released MiniMax H3, an open video generation model that matches ByteDance's Seedance 2.0 in cloud performance. Its Omni Reference feature accepts up to nine images, three videos, and three audio tracks, and it runs on a local PC with an RTX 4090. ASCII.jp tested it with live-action and anime-style prompts.
Aug 9
Chinese AI firm MiniMax announced MiniMax H3, a video generation model that creates up to 15 seconds of audio-synced video from text, images, or voice. It supports 768p or 2K resolution and accepts multiple media inputs at once. In a blind test by Artificial Analysis, H3 ranked second for text-to-video, behind Gemini Omni Flash and ahead of Seedance 2.0. The model is available via API now, with open weights planned soon.