Artificials
Compare models
All makers

AI model maker

MiniMax

17 models on Artificials, 5 with published test results and 15 with open weights you can download.

RSS feed
Best on the capability index146.9MiniMax M3, #68 of 274
Best rankTop 5%MiniMax-M2.5 (high) on SWE-bench Verified, bash only
Newest releaseMiniMax-M3.1-Flash-PreviewSep 27, 2026
Models175 tested, 15 with open weights

Where MiniMax stands

Its best model’s rank among all models tested on each skill area’s main test, the same tests model pages use. A longer bar means a better rank.

  • OverallMiniMax M3 on Epoch Capabilities Index
    Top 25%
  • CodingMiniMax-M3 on LiveBench coding
    #61 of 63
  • MathMiniMax-M3 on Mock AIME 2024–2025
    Top 46%
  • AgentsMiniMax-M3 on LiveBench agentic coding
    #58 of 63
  • LanguageMiniMax-M3 on LiveBench language
    #42 of 63
  • Data analysisMiniMax-M3 on LiveBench data analysis
    #32 of 63
  • ReasoningMiniMax-M3 on GPQA Diamond
    Top 14%

On the capability index

Its 3 models on the Epoch Capabilities Index, each with its rank among all 274 models. The top score is 167.3, held by Claude Opus 5.5 from Anthropic.

Every test

Its best result on each published test, and that result’s rank among every model tested. Scores from different tests are never combined.

TestBest modelScoreRank
Epoch Capabilities IndexMiniMax M3146.9#68 of 274
SWE-bench Verified, bash onlyMiniMax-M2.5 (high)75.8%#2 of 42
GPQA DiamondMiniMax-M390.9%#31 of 222
Arena codingMiniMax-M31494#74 of 384
Arena instruction followingMiniMax-M31434#76 of 389
Arena hard promptsMiniMax-M31463#77 of 389
Arena textMiniMax-M31440#83 of 389
Arena creative writingMiniMax-M31406#84 of 387
Arena mathMiniMax-M31432#90 of 372
Arena WebDevMiniMax-M31482#51 of 123
Mock AIME 2024–2025MiniMax-M371.1%#91 of 198
LiveBench data analysisMiniMax-M376.2%#32 of 63
Chess puzzlesMiniMax-M314.0%#73 of 141
LiveBench languageMiniMax-M376.8%#42 of 63
LiveBench agentic codingMiniMax-M340.7%#58 of 63
LiveBench overallMiniMax-M367.3%#58 of 63
LiveBench reasoningMiniMax-M374.5%#58 of 63
LiveBench instructionsMiniMax-M357.5%#59 of 63
LiveBench codingMiniMax-M368.2%#61 of 63
LiveBench mathematicsMiniMax-M377.0%#62 of 63

Every model

Newest first. The price is per 1M tokens, blended three input tokens to one output token: MiniMax’s own listing, or else the middle price across hosts.

ModelReleasedContextPriceCapability index
MiniMax-M3.1-Flash-PreviewSep 27, 20261M tokensNo paid listingNot on the index
MiniMax-M3Jun 1, 20261M tokens$0.525146.9
MiniMax LatestMay 3, 2026512K tokens$0.525Not on the index
MiniMax-M2.7Mar 18, 2026204.8K tokens$0.525145.8
MiniMax-M2.7-highspeedMar 18, 2026204.8K tokens$1.05Not on the index
MiniMax-M2.5 LightningFeb 13, 2026204.8K tokens$0.825Not on the index
MiniMax-M2.5-highspeedFeb 13, 2026204.8K tokens$1.05Not on the index
MiniMax-M2.5Feb 12, 2026204.8K tokens$0.525146.7
MiniMax-M2 HerJan 23, 202665.5K tokens$0.525Not on the index
MiniMax M2.1 LightningDec 23, 2025204.8K tokens$0.21Not on the index
MiniMax-M2.1Dec 23, 2025204.8K tokens$0.525Not on the index
MiniMax-M2Oct 27, 2025204.8K tokens$0.525Not on the index
MiniMax M2.5Mar 1, 2025204.8K tokens$0.44Not on the index
MiniMax Text 01Jan 15, 20251M tokens$0.425Not on the index
MiniMax M1Jan 8, 20251M tokens$0.438Not on the index

Models, prices, context windows and release dates from models.dev (MIT). Test results from LiveBench (Apache-2.0), Epoch AI (CC BY 4.0), Arena (CC BY 4.0) and SWE-bench (CC BY-NC 4.0). A model counts here when MiniMax sells it, several hosts list it or a test covers it. Each figure is the source’s own; nothing is combined into a new score.