modelbenchmark.io

MiniMax-M3

MiniMax · minimax-m3

Compare

MiniMax multimodal model for long-context coding, perception, and agent planning

Specification

minimax listing

Context
1M
Max output
524K
Released
2026-06-01
Knowledge cutoff
2025-01
Retires
2026-11-06
Open weights
yes
Input
text, image, video
Output
text

Price

US dollars per million tokens · minimax list price

Input
$0.30
Output
$1.20
Cache read
$0.06
Cache write
$0.90

Batch and priority prices

US dollars per million tokens · per listing

PriceListing$/M
Priority inputfireworks_ai$0.45
Priority outputfireworks_ai$1.80

Long-context price tiers

the rate above each context threshold

AbovePrice$/M
512KCache read$0.12
512KInput$0.60
512KOutput$2.40

Quality

5 benchmarks · 2 sources

Composite
47th
percentile of 344 scored models
Rank
183 / 344
±0.11 sd
Evidence
5 × 2
3 or more benchmarks, from 2 or more sources
Effort range
—
one configuration only
reasoning51st3/4 bench
math75th1/7 bench
general9th1/1 bench

Every score

one row per source and configuration — nothing averaged away

BenchmarkScoreConfigurationSourceRun
GPQA diamond90.9 ±2.0defaultEpoch AI2026-08-10
OTIS Mock AIME 2024-202571.1 ±6.8defaultEpoch AI2026-08-10
Chess Puzzles13.0 ±3.4defaultEpoch AI2026-10-13
Mystery Game Puzzles7.0 ±2.6defaultEpoch AI2026-08-27
LiveBench67.5default · 23 runsLiveBench2026-06-25

The leading figure for a benchmark is the median across its configurations, so one heroic high-effort run cannot set the number.

Available from 75 hosts

HostIn $/MOut $/MCache rdCache wrContextOutputRetires
minimaxvendor$0.30$1.20$0.06—1M524K—
minimax-cn-coding-planfreefreefreefree1M524K—
minimax-coding-planfreefreefreefree1M524K—
scnet-token-planfreefreefree—1M512K—
volcengine-coding-planfreefreefree—1M512K—
vultr$0.20$0.90——524K512K—
empiriolabs$0.225$0.90$0.045—1M524K—
ambient · minimax$0.23$0.96$0.05free131K132K—
nano-gpt · minimax$0.23$0.96$0.05—512K80K—
nano-gpt · thinking$0.23$0.96$0.05—512K80K—
wandb · minimaxai$0.23$0.96$0.05—262K262K—
wandb · wandb$0.23$0.96$0.05—262K262K—
vancine$0.24$0.96$0.048—1M512K—
boundless$0.28$1.10$0.056—1M512K—
deepinfra$0.28$1.10$0.056—524K524K—
deepinfra · minimaxai$0.28$1.10$0.056—524K512K—
lilac · minimaxai$0.28$1.10$0.05—1M1M—
llmgateway$0.28$1.10$0.05—1M512K—
llmgateway-providers · consensusprotocol$0.28$1.10$0.05—1M512K—
aihubmix$0.288$1.152——1M512K—
aihubmix · aihubmix$0.288$1.152——1M524K—
abacus · minimaxai$0.30$1.20——1M131K—
cline-pass · cline-pass$0.30$1.20$0.06—1M512K—
commandcode · minimaxai$0.30$1.20$0.06—1M512K—
cortecs$0.30$1.20$0.06—1M1M—
edenai · minimax$0.30$1.20$0.06—524K512K—
fireworks_ai$0.30$1.20$0.06—512K512K—
fireworks-ai · accounts$0.30$1.20$0.06—512K512K—
huggingface · minimaxai$0.30$1.20——524K512K—
inco$0.30$1.20$0.06—1M512K—
iteracompute · minimax$0.30$1.19$0.05—1M524K—
jalapeno$0.30$1.20$0.06—524K512K—
kilo · minimax$0.30$1.20$0.06—524K512K—
llmgateway-providers$0.30$1.20$0.06—524K131K—
llmgateway-providers · tencent$0.30$1.20$0.06—1M131K—
merge-gateway · minimax$0.30$1.20$0.06—1M128K—
minimax · minimax$0.30$1.20$0.06—1M128K—
modular · minimax$0.30$1.20$0.06—1M512K—
nebius$0.30$1.20——1M1M—
nebius · minimaxai$0.30$1.20$0.06—1M1M—
novita$0.30$1.20$0.06—1M131K—
novita-ai · minimax$0.30$1.20$0.06—1M131K—
opencode$0.30$1.20$0.06—512K128K—
opencode_go · opencode-go$0.30$1.20$0.06—1M131K—
opencode-go$0.30$1.20$0.06—1M131K—
openrouter · minimax$0.30$1.20$0.06—1M512K—
orcarouter · minimax$0.30$1.20$0.06—1M512K—
requesty$0.30$1.20$0.06—1M128K—
siliconflow · minimaxai$0.30$1.20$0.06—1M131K—
standardcompute · minimax$0.30$1.20$0.06—1M50K—
tempr · minimax$0.30$1.20$0.06—1M512K—
tencent · tencent$0.30$1.20$0.06free1M——
together_ai$0.30$1.20$0.06—524K524K2026-11-06
togetherai · minimaxai$0.30$1.20$0.06—524K524K—
venice$0.30$1.20$0.06—524K66K—
vercel · minimax$0.30$1.20$0.06—512K512K—
zenmux · minimax$0.30$1.20$0.06—1M512K—
minimax-cn$0.3137$1.2549$0.0627—1M524K—
hyper$0.3266$1.3066$0.0642—512K512K—
crossmodel · minimax$0.33$1.32$0.066$0.331M512K—
wafer.ai$0.33$1.32$0.07free1M512K—
wally$0.33$1.32$0.066—524K131K—
kenari$0.36$1.4286$0.0686—1M512K—
inferenceapis · inferenceapis$0.39$1.56$0.08—524K524K—
kyma · kyma$0.3956$1.582$0.0756—1M33K—
tensorx · minimax$0.40$2.00$0.10—1M131K—
gmicloud · minimaxai$0.60$2.40$0.12—1M512K—
inco · fast$0.60$2.40$0.12—1M512K—
llmgateway-providers · minimax$0.60$2.40$0.12—512K131K—
ofox · minimax$0.60$2.40$0.12—1M512K—
ollama-cloud$0.60$2.40$0.12—512K131K—
opper$0.60$2.40$0.12—1M131K—
sambanova$0.60$2.40$0.06—1M1M—
streamlake · streamlake$0.60$2.40$0.12—1M512K—
302ai$0.72$2.88$0.144$0.901M512K—

Price history

append-only observations · a listing writes a row only when its price moves

DateHostField$/M
2026-09-02cortecsCache read$0.099
2026-09-26cortecsCache read$0.06
2026-09-02kilo · minimaxCache read$0.06
2026-10-23kilo · minimaxCache read$0.0448
2026-10-28kilo · minimaxCache read$0.06
2026-09-02llmgatewayCache read$0.06
2026-10-15llmgatewayCache read$0.05
2026-09-02minimax-cnCache read$0.06
2026-10-20minimax-cnCache read$0.0627
2026-09-02nano-gpt · minimaxCache read$0.06
2026-10-19nano-gpt · minimaxCache read$0.05
2026-09-02nano-gpt · thinkingCache read$0.06
2026-10-19nano-gpt · thinkingCache read$0.05
2026-09-02cortecsInput$0.395
2026-09-26cortecsInput$0.30
2026-09-02kenariInputfree
2026-10-22kenariInput$0.36
2026-09-02kilo · minimaxInput$0.30
2026-10-23kilo · minimaxInput$0.224
2026-10-28kilo · minimaxInput$0.30
2026-09-02llmgatewayInput$0.30
2026-10-15llmgatewayInput$0.28
2026-09-02minimax-cnInput$0.30
2026-10-20minimax-cnInput$0.3137
2026-09-02nano-gpt · minimaxInput$0.30
2026-10-19nano-gpt · minimaxInput$0.23
2026-09-02nano-gpt · thinkingInput$0.30
2026-10-19nano-gpt · thinkingInput$0.23
2026-09-02openrouter · freeInputfree
2026-09-06openrouter · freeInputfree
2026-09-08openrouter · freeInputfree
2026-09-02zenmux · minimaxInput$0.60
2026-10-22zenmux · minimaxInput$0.30
2026-09-02cortecsOutput$1.977
2026-09-26cortecsOutput$1.20
2026-09-02kenariOutputfree
2026-10-22kenariOutput$1.4286
2026-09-02kilo · minimaxOutput$1.20
2026-10-23kilo · minimaxOutput$0.88
2026-10-28kilo · minimaxOutput$1.20
2026-09-02llmgatewayOutput$1.20
2026-10-15llmgatewayOutput$1.10
2026-09-02minimax-cnOutput$1.20
2026-10-20minimax-cnOutput$1.2549
2026-09-02nano-gpt · minimaxOutput$1.20
2026-10-19nano-gpt · minimaxOutput$0.96
2026-09-02nano-gpt · thinkingOutput$1.20
2026-10-19nano-gpt · thinkingOutput$0.96
2026-09-02openrouter · freeOutputfree
2026-09-06openrouter · freeOutputfree
2026-09-08openrouter · freeOutputfree
2026-09-02zenmux · minimaxOutput$2.40
2026-10-22zenmux · minimaxOutput$1.20
2026-09-17iteracompute · minimaxCache read$0.08
2026-10-16iteracompute · minimaxCache read$0.045
2026-10-22iteracompute · minimaxCache read$0.05
2026-09-17iteracompute · minimaxInput$0.29
2026-10-16iteracompute · minimaxInput$0.299
2026-10-22iteracompute · minimaxInput$0.30
2026-09-17iteracompute · minimaxOutput$1.20
2026-10-16iteracompute · minimaxOutput$1.19
2026-10-22nebius · minimaxaiCache read$0.30
2026-10-24nebius · minimaxaiCache read$0.06

More from MiniMax

most-hosted first