AI Model Radar
中文 EN

MiniMax-M3

MiniMax
Model overview

MiniMax-M3

An open-weight MiniMax model for million-token context, coding, and multimodal agent workflows.

general LLMreasoningcodingmultimodalvision-languagelong contextagentchatcode generationsoftware engineering agentlocal inferenceenterprise access
External observationsCross-model riskCompareCompanyRankingsPublic signals
Model risk

Current observations

1current adverse observations
Negative / riskAPI / serviceContext availabilityOfficial · Current risk

API prices for 512K-to-1M context are double the rates at or below 512K

MiniMax's enterprise API pricing page shows that input, cache, and output rates for the 512K-to-1M input range are twice the rates at or below 512K. The current page no longer supports the earlier limited-availability claim.

MiniMax enterprise API token pricingHigh confidenceObserved 2026-06-01Source as of 2026-07-15 Action

Budget million-context workloads against the separate tier and regression-test behavior at the 512K boundary.

Open source
Representative version

MiniMax-M3

CompanyMiniMax
Release2026-06-01
Release typeModel release
Parameters428B / 23B active
ArchitectureMixture-of-experts transformer with MiniMax Sparse Attention
Context1,000,000
Input modalitiestext / image / video
Output modalitiestext / code / tool_call
Tool use / function callingConfirmed: Tool use, Function calling
Structured output / reasoningConfirmed: Reasoning
Vision / audio / videoConfirmed: Vision, Video
Open weights / LicenseWeights available / MiniMax Community License
License sourceminimax-m3-checkpoint-2026-06-12 · 2026-08-04
Vendor APIAvailable
MiniMax M3 release articleminimax-m3-current · 2026-06-01
OpenRouterOpenRouter listed
PricingOfficial · 2026-06-01: $0.3 / $1.2
OpenRouter · 2026-08-11: $0.3 / $1.2
input / output, per 1M tokens
BenchmarkNo reliable public source
Public signals1 community discussion / 1 activity snapshot
Use casesLocal multimodal inference / coding and review / Long-context agents / research evaluation
Claim evidenceRelease eventMiniMax M3 release articleMiniMax-M3 · 2026-06-01
Model specifications / Weight checkpointMiniMax-M3 model cardMiniMax-M3 · MiniMax-M3 · MiniMax-M3
OpenRouter context / capabilities / pricingOpenRouter public models APIminimax/minimax-m3 · 2026-08-11T06:28:09.895Z
OpenRouter

Pricing / context / availability

OpenRouterOpenRouter listed
OpenRouter listingminimax/minimax-m3
Mapped versionMiniMax-M3
minimax-m3-current
PricingOfficial · 2026-06-01: $0.3 / $1.2
OpenRouter · 2026-08-11: $0.3 / $1.2
input / output, per 1M tokens
Context1048576
Speed/LatencyNo reliable public source
Snapshot time2026-08-11T06:28:09.895Z
SourceView source
Source-bounded public ranking

OpenRouter weekly token-usage ranking

SourceModel / versionMetricRankValueCheckedLink
OpenRouterMiniMax: MiniMax M3 0531
minimax-m3-current		Weekly token usage on OpenRouter public rankings API111,686,799,743,979 tokens2026-08-11View source

OpenRouter weekly prompt-plus-completion token usage only; this is not a model-quality or whole-market ranking.

Verified limitations

Usage boundaries

  • The 512K-1M context tier doubles the up-to-512K input and output rates.
  • The June 1 release made M3 available through MiniMax products and API; public model weights followed on June 12, 2026.
Public activity counters

Hugging Face model-card activity snapshot

PlatformEntryLast-month downloadsCurrent likesCurrent SpacesTimeSource
huggingface
Current HF activity snapshot		MiniMaxAI/MiniMax-M3156,7831,44941View source

Downloads are Hugging Face's trailing-month counter; likes and Spaces are snapshots at collection time.

Vertical comparison

Current available versions

VersionReleaseParametersContextInputOutputOpen weights / LicenseVendor APISource tier
MiniMax-M32026-06-01Not disclosed1,000,000text / image / videotext / code / tool_callNo open weights currently / weight license not disclosedAvailableOfficial source
MiniMax-M32026-06-12428B / 23B active1,000,000text / image / videotext / code / tool_callWeights available / MiniMax Community LicenseNot disclosedOfficial source
Missing public information

Public information not yet available

Parameter scale not disclosedNo reliable benchmark sourceNo linkable model-specific reviewSpeed/latency not disclosed