Model Releases

MiniMax M2.7: early echoes of self-evolution

Original title:MiniMax M2.7: Early Echoes of Self-Evolution

AI summary

M2.7 is the first model in the M2 series to deeply participate in its own evolution. It can build complex agent frameworks and complete refined productivity tasks, especially in software engineering, scoring 56.22% on the SWE-Pro benchmark, close to Opus's best. Its office software handling leads among open-source models, with an ELO score of 1495 on GDPval-AA. M2.7 maintains a 97% skill adherence rate, handling more than 40 complex skills each exceeding 2,000 tokens. Through an internal research agent framework, the model achieves an autonomous 'analyze-modify-evaluate' iterative optimization loop, improving performance in internal evaluations.

Key Takeaways

  • 7 is the first model in the M2 series to deeply participate in its own evolution.
  • It can build complex agent frameworks and complete refined productivity tasks,especially in…
  • 22% on the SWE-Pro benchmark,close to Opus's best.
  • Its office software handling leads among open-source models,with an ELO score of 1495 on GDPval-AA.
  • 7 maintains a 97% skill adherence rate,handling more than 40 complex skills each exceeding 2,000…

Quick Facts

Published
Updated
Source
MiniMax:Blog(网页)
Category
Model Releases
AI score
61
Read the original Read the full article at the original source
Why it matters

MiniMax M2.7 让模型参与自身进化,在 SWE-Pro 和 VIBE-Pro 上接近 Opus 水平,Agent Teams 设计也值得看,但整体仍是追赶者姿态。

Related entities

Source & copyright

This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.

Original article · AIHOT entry

How to Cite

When citing this article, keep the title, date, and permalink, and credit Orange News as the source.

Orange News: MiniMax M2.7: early echoes of self-evolution. March 18, 2026. https://new.nat6.net/en/news/cmpunu0g50005slbro8j0g9i9/

View all →