Skip to content
← Models
Open weightsMiniMax/10 trustVerified
MiniMax M2.1 logo

MiniMax M2.1

Open-weight agentic coding model, 10x cheaper than Claude at launch.

Visit official page →

About

MiniMax M2.1 is MiniMax's updated open-weight coding model, released December 23, 2025, the same day as GLM-4.7. Built for agentic coding and tool-calling workflows, it pairs high inference efficiency with strong multilingual software engineering performance, scoring 72.5% on the SWE-multilingual benchmark and 88.6% aggregate on MiniMax's own VIBE-bench for full-stack app development.

Weights are hosted on Hugging Face for local deployment with vLLM and SGLang, and the model is served through the MiniMax Open Platform API, priced around $0.30 per million input tokens at launch. Vendor benchmark figures are self-reported and should be treated as such. It also powers the MiniMax Agent platform for end-to-end office and development tasks.

Specs

Context window

Pricing

Free quota

API access

Weights

Open weights

Released

December 23, 2025

Benchmarks

Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.

Use cases

coding-agentstool-callingmultilingual-devself-hosted-llm

Trust score — 8/10

/10 trustVerified

Scored 8/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 89; Trending score 82/100 from researcher assessment.

More in the minimax-m2 family

Sources