Paper Splits Conformal Reliability From Frequency Calibration After a Break
A Columbia University paper on arXiv separates two audits of conformal prediction, a method relevant to pay-per-token vendors such as MiniMax.
MiniMax news
Every change we have logged at MiniMax, newest first, each with the page MiniMax published as the receipt and a plain statement of what it means for you.
Three items. Back to 8 Jul 2026. Research. Updated 21 Oct 2026.
What is the latest MiniMax news?
The newest change logged at MiniMax is Paper Splits Conformal Reliability From Frequency Calibration After a Break, dated 3 Oct 2026. The cheapest MiniMax model this ledger prices costs $0.30 per million input tokens, read from platform.minimax.io/docs/guides/pricing-paygo on 2 Nov 2026.
Written from the newest item on this page and from the ledger rows behind it. Newest event .
Where MiniMax stands, and what we hold on itevery company
Paper Proposes Sparse Priors To Cut Sample Needs In Distribution Learning
A new arXiv paper argues that sparse priors can sharply lower the number of samples needed to learn a distribution, a bound relevant to model vendors such as MiniMax.
Researchers Publish Bias-Variance Theory For Guided Reasoning Training
A paper relevant to reasoning methods used across the field, including by vendors such as MiniMax, sets out a bias-variance theory for guidance-augmented reinforcement learning.
Items here are written from MiniMax's own announcements, documentation and price pages, never from another publisher, and every item carries both the day it happened and the day we published it. How this page is made. Published by Kawsar. RSS feed for MiniMax alone.