Coding
A preview coding model was reported in late September; evaluate cautiously.
A multimodal lab spanning text, speech, music and video models.
The short version
Overview
MiniMax builds language models alongside speech, music and video generation. Its M-series models are multimodal, and its Hailuo models generate video.
It is notable for covering many media types from one company, with a coding tool built around its newer models.
Current models. Public reports describe MiniMax M3 (June 2026), an M3.1 Flash Preview (late September 2026) and an H3 video model (July 2026). Sources vary on names and dates.
Where it is strong
Scores are Halden’s editorial view of typical fit on a 0–5 scale. They are not benchmark results and will differ by task, model version and plan: test on your own work.
Coding, analytics, documents, research
A preview coding model was reported in late September; evaluate cautiously.
Adequate; not its focus.
Capable multimodal drafting.
Usable; verify claims.
Practical tips
What has changed recently
Described as a natively multimodal sparse mixture-of-experts model with up to a one-million-token context.
Source: DataLearnerReports say a lighter preview coding model arrived in late September.
Source: Hermes AICompiled on October 9, 2026 from public reporting, much of it secondary. Dates, names and specifications can change, so confirm on MiniMax’s own site before you rely on them.
Go deeper
Keep exploring
FAQ
Multimodal generation across text, speech, music and video.
MiniMax's video generation brand.
Some MiniMax models have been released with open weights; check each one.
A next step
Run a short pilot with the people who will use MiniMax, measure the result, and decide with data, not hype.
Free assessments
Five short assessments, an instant profile and three practical next steps.
Stay up to date with practical AI ideas and insights
Privacy Policy