The Model Index
Up to four models

MiniMax-M2.5 vs MiniMax-M2.1

On the index (GPT-4 = 100), MiniMax-M2.5 scores 144 and MiniMax-M2.1 135. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.

Next to MiniMax-M2.5
Presets
MiniMax-M2.5
144128–160
WeightsOpen
ReleasedFeb 2026
Price in / out$0.29 / $1.2
Context198K
Core tests taken1 of 9
MiniMax-M2.1
135117–153
WeightsOpen
ReleasedDec 2025
Price in / out–
Context–
Core tests taken0 of 9

MiniMax-M2.5 and MiniMax-M2.1 are too close to call (8 points apart, within the uncertainty).

Test by testthird-party results

MiniMax-M2.5MiniMax-M2.1
0%25%50%75%100%REASONING & KNOWLEDGEGPQA Diamond (Vals AI)GPQA Diamond (Vals AI)MMLU ProMMLU ProCODINGTerminal-Bench 2.0 (Terminus 2 runs)Terminal-Bench 2.0 (Terminus 2 runs)LiveCodeBenchLiveCodeBenchSWE-benchSWE-benchPROFESSIONAL WORKCorpFin v2CorpFin v2LegalBenchLegalBenchTaxEval v2TaxEval v2
Every result as a table
TestMiniMax-M2.5MiniMax-M2.1
Reasoning & knowledge
GPQA Diamond (Vals AI)82.1%78.5%
MMLU Pro80.1%87.1%
Coding
Terminal-Bench 2.0 (Terminus 2 runs)42.3%29.2%
LiveCodeBench79.2%81.8%
SWE-bench74.2%74.8%
Professional work
CorpFin v259.6%58.9%
LegalBench80.0%75.4%
TaxEval v268.2%66.3%

Tests at least two of these models have taken. Bold rows are the core tests.

This page as a chartDownload PNG