Noometry

MiniMax, open weights

MiniMax-M3

MiniMax-M3 by MiniMax ranks 85th of 354 ranked models on the Noometry Index as of October 2026, with a score of 43.8. Its strongest category is knowledge, where it ranks 35th. API pricing starts at $0.30 per million input tokens and $1.20 per million output tokens, with a 1M-token context window.

Last verified

Noometry rank
#85 of 354
Index score
43.8
Evidence
Confirmed 41 results
Provider
MiniMax
Released
June 1, 2026
Weights
Open weights
Reasoning
Yes
Context window
1M
Max output
512K
Input price
$0.30 / M
Output price
$1.20 / M
Blended price
$0.52 / M
Output speed
Not measured
Value
#73 of 219
Knowledge cutoff
Unknown
Input
text, image, video
Hugging Face
MiniMaxAI/Minimax-M3

Category scores

Each category score combines every public result we have in that category.

  1. Coding 41.8
  2. Agentic & Tool Use 22.6
  3. Reasoning 30.1
  4. Math 40.0
  5. Knowledge 58.4
  6. Multimodal 40.2
  7. Multilingual 53.0
  8. Instruction Following 75.5
  9. Long Context 44.2
  10. Writing & Preference 62.1
CategoryScoreRankResults
Coding41.8#1184
Agentic & Tool Use22.6#1303
Reasoning30.1#879
Math40.0#953
Knowledge58.4#352
Multimodal40.2#511
Multilingual53.0#751
Instruction Following75.5#621
Long Context44.2#721
Writing & Preference62.1#834

Strengths and weaknesses

Categories where MiniMax-M3 places highest and lowest among the models ranked in each, with its score against that category's median.

Strongest categories

CategoryScorevs medianRank
Knowledge58.4+21.1#35 of 314, top 12%
Instruction Following75.5+4.2#62 of 305, top 21%
Long Context44.2+3.2#72 of 296, top 25%

Weakest categories

CategoryScorevs medianRank
Agentic & Tool Use22.6−7.8#130 of 154, top 85%
Multimodal40.2+1.6#51 of 128, top 40%
Coding41.8+3.1#118 of 340, top 35%

Closest competitors

The models ranked just above and below MiniMax-M3. When scores are this close, price and speed are often the better way to choose.

ModelRankScoreBlended $/MSpeed
Claude Sonnet 4.5#8144.1$685Compare
Chatgpt 4o Latest 20250326#8243.8—21Compare
ERNIE 5.1#8343.8——Compare
GLM-5V-Turbo#8443.8$1.90—Compare
Grok 4.3#8643.8$1.56—Compare
Qwen3 Max#8743.7$2.4048Compare
MiMo-V2-Omni#8843.6$0.18—Compare
Kimi K2.5 Instant#8943.6——Compare

Sponsored placements are available on pages like this one. Advertise on Noometry

Benchmark results

Every published result we track, with its source. Bold rows are the ones used for ranking; where several exist we prefer independent runs over self-reported numbers.

Coding

BenchmarkScorePositionSettingSourceDate
FrontierCode14.7%#33 of 37, top 90%Epoch AI
LMArena WebDev1482#51 of 113, top 46%LMArena2026-10-08
SciCode47.1%#52 of 121, top 43%Epoch AI
LMArena Coding1469#58 of 294, top 20%LMArena2026-10-08
ALE-Bench640.02#76 of 105, top 73%Epoch AI

Agentic & Tool Use

BenchmarkScorePositionSettingSourceDate
APEX-Agents37.7%#38 of 49, top 78%Epoch AI
OSWorld 2.04.6%#8 of 9, top 89%Epoch AI
GBAEval0.9%#16 of 23, top 70%Epoch AI
Vending-Bench 22,158#43 of 60, top 72%Epoch AI

Reasoning

BenchmarkScorePositionSettingSourceDate
SimpleBench45.8%#48 of 77, top 63%Epoch AI
NYT Connections (extended)65.1%#54 of 91, top 60%Lech Mazur benchmarks
CritPt3.7%#62 of 134, top 47%Epoch AI
Chess Puzzles14%#71 of 129, top 56%Epoch AI2026-08-10
Chess Puzzles4%noneEpoch AI2026-08-10
LMArena Hard Prompts1447#70 of 297, top 24%LMArena2026-10-08
Mystery Game Puzzles7%Epoch AI2026-08-27
Mystery Game Puzzles8%#65 of 74, top 88%noneEpoch AI2026-08-27
DTBench78.9%#78 of 151, top 52%Epoch AI
LMCA33.7%#67 of 125, top 54%Epoch AI
Surface Evolver Bench55%#13 of 25, top 52%Epoch AI
Epoch Capabilities Index146.95#66 of 213, top 31%Epoch AI2026-06-01
ForecastBench61.4#12 of 72, top 17%Epoch AI

Math

BenchmarkScorePositionSettingSourceDate
OTIS Mock AIME 2024-202571.1%#91 of 173, top 53%Epoch AI2026-08-10
OTIS Mock AIME 2024-202526.7%noneEpoch AI2026-08-10
ProofBench18%#52 of 77, top 68%Epoch AI
LMArena Math1429#87 of 285, top 31%LMArena2026-10-08

Knowledge

BenchmarkScorePositionSettingSourceDate
GPQA Diamond90.9%#30 of 186, top 17%Epoch AI2026-08-10
GPQA Diamond81.3%noneEpoch AI2026-08-10
LMArena Expert1461#61 of 273, top 23%LMArena2026-10-08

Multimodal

BenchmarkScorePositionSettingSourceDate
LMArena Vision1253#52 of 122, top 43%LMArena2026-10-09
LMArena Document1435#27 of 38, top 72%LMArena2026-09-13

Multilingual

BenchmarkScorePositionSettingSourceDate
LMArena Non-English1420#75 of 297, top 26%LMArena2026-10-08
LMArena Chinese1463#85 of 285, top 30%LMArena2026-10-08
LMArena French1447#78 of 223, top 35%LMArena2026-10-08
LMArena German1426#70 of 231, top 31%LMArena2026-10-08
LMArena Japanese1381#77 of 211, top 37%LMArena2026-10-08
LMArena Korean1372#84 of 213, top 40%LMArena2026-10-08
LMArena Russian1428#71 of 283, top 26%LMArena2026-10-08
LMArena Spanish1432#80 of 226, top 36%LMArena2026-10-08

Instruction Following

BenchmarkScorePositionSettingSourceDate
LMArena Instruction Following1433#58 of 298, top 20%LMArena2026-10-08

Long Context

BenchmarkScorePositionSettingSourceDate
LMArena Longer Query1445#59 of 291, top 21%LMArena2026-10-08

Writing & Preference

BenchmarkScorePositionSettingSourceDate
LMArena Text1433#75 of 297, top 26%LMArena2026-10-08
LMArena Creative Writing1404#69 of 295, top 24%LMArena2026-10-08
EQ-Bench 41150#20 of 28, top 72%EQ-Bench
LMArena Multi-Turn1442#69 of 295, top 24%LMArena2026-10-08

API pricing by provider

RouteInput $/MOutput $/MCached input $/MChecked
deepinfra$0.28$1.10$0.0562026-10-10
fireworks$0.30$1.20$0.062026-10-10
minimax$0.30$1.20$0.062026-10-10
openrouter$0.30$1.20$0.062026-10-10
together$0.30$1.20$0.062026-10-10

All MiniMax API prices → Estimate your cost →

Compare MiniMax-M3

Other MiniMax models

Frequently asked questions

How good is MiniMax-M3?

MiniMax-M3 by MiniMax ranks 85th of 354 ranked models on the Noometry Index as of October 2026, with a score of 43.8. Its strongest category is knowledge, where it ranks 35th. API pricing starts at $0.30 per million input tokens and $1.20 per million output tokens, with a 1M-token context window.

How much does MiniMax-M3 cost?

MiniMax-M3 costs $0.30 per million input tokens and $1.20 per million output tokens on MiniMax's own API, with cached input at $0.06.

What is MiniMax-M3's context window?

MiniMax-M3 accepts up to 1M tokens of input and can write up to 512K tokens in one response.

Is MiniMax-M3 open source?

Yes. MiniMax-M3's weights are downloadable from Hugging Face (MiniMaxAI/Minimax-M3); check the license for commercial terms.

What are MiniMax-M3's strengths and weaknesses?

Relative to other ranked models, MiniMax-M3 places best in knowledge, instruction following, long context and lowest in agentic & tool use, multimodal, coding.

What is MiniMax-M3 best at?

Its best category is knowledge, where it ranks 35th on Noometry.