↓ Skip to main content
9 – 17 UHR +49 8031 3508270 LUITPOLDSTR. 9, 83022 ROSENHEIM
DE / EN
LLM MiniMax AI China

MiniMax-M3

MiniMax-M3.1-Flash and MiniMax-M3: open-weight models with 1M context for agentic coding. API pricing, licence, Bedrock and Foundry availability, GDPR self-hosting.

License MiniMax Community License (M3.1 Flash, M3); Non-Commercial License (M2.7); Modified MIT (M2, M2.1, M2.5)
GDPR Hosting Available
Context 1M Tokens
Modality Text, Image, Video → Text

Versions

Overview of available model variants

ModelReleaseEUStrengthsWeaknessesStatus
MiniMax-M3.1-Flash Recommended
2026-10-19
Around 428B parameters, 23B active (MoE), 1M-token context, text, image and video input MiniMax Sparse Attention in every layer, quantization-aware training and DSpark speculative decoding – considerably smaller weight files than M3 Per MiniMax: DeepSWE v1.1 73.7%, BankerToolBench 76.2, OSWorld 2.0 61.7 and CyberGym 82.2 Thinking depth configurable in five levels (low, medium, high, xhigh, max; default max) Per MiniMax around 70% lower usage cost on agentic tasks than its predecessor
MiniMax Community License instead of MIT: attribution and a one-time notice to MiniMax, prior written authorisation above US$20M annual revenue Benchmarks mostly measured by MiniMax itself No offering with EU data residency from MiniMax itself
Current
MiniMax-M3
2026-06-01
1M-token context window (per MiniMax, at least 512k guaranteed in the API) Native multimodality (image and video understanding) MSA architecture (MiniMax Sparse Attention) – per MiniMax 1/20 of the predecessor's per-token compute at 1M context Per MiniMax more than 9x faster prefill and more than 15x faster decoding at 1M context SWE-Bench Pro 59.0% and Terminal-Bench 2.1 66.0% (per MiniMax) Selectable reasoning: enabled, adaptive or disabled Open weights on Hugging Face (MiniMaxAI/MiniMax-M3, ~428B parameters, ~23B active)
Overtaken by M3.1 Flash on price and efficiency Self-hosting still experimental per MiniMax (SGLang development image), reference hardware 8x B200 with MXFP8 weights Licensed under the MiniMax Community License, not MIT: commercial use requires attribution ('Built with MiniMax M3') and, above US$20M annual revenue, prior written authorisation from MiniMax; below that threshold a one-time notice suffices
Current
MiniMax-H3 (video, open weights H3-Base)
2026-07-31
Text, image, video and audio as a shared context, output as video with native stereo sound (32 kHz), 4 to 15 seconds at 24 fps Open H3-Base weights (33B parameters, dense single-stream transformer) for 768p; 2K only in combination with the hosted modules H3-Context-IR and H3-Regenerate-2K Per the model card 11 languages with stable dialogue output
Licence explicitly excludes the EU – not deployable locally for companies in the DACH region without a special licence International API with data stored in the US, no EU residency Additional written authorization required above USD 20 million annual revenue; attribution requirement 'Powered by MiniMax H3' Open weights deliver 768p only, full 2K quality remains tied to the hosted API; deployment example with four GPUs
Current
Older, deprecated and retired versions (3)

Legacy and deprecated: superseded by newer models; for many versions the provider has already announced a shutdown. Retired: no longer available. Dates and successors are in the model lifecycle section.

ModelReleaseEUStrengthsWeaknessesStatus
MiniMax-M2.7
2026-03-18
Sparse Mixture-of-Experts (MoE) architecture Designed for agent teams and complex coding workflows Per MiniMax 56.22% on SWE-Pro, 57.0% on Terminal Bench 2 and a 66.6% medal rate on MLE-Bench Lite Native multi-agent collaboration
Superseded by M3 and M3.1 Flash; the MiniMax API ends on 18 November 2026 Not MIT-licensed: per the LICENSE file on Hugging Face (MiniMaxAI/MiniMax-M2.7), any commercial use is prohibited without prior written authorisation from MiniMax – stricter than the MiniMax Community License of M3 and M3.1 Flash, which only requires a notice below US$20M annual revenue No native EU cloud availability
Deprecated Shutdown: 18 November 2026 (API)
MiniMax-M2.5
February 2026
Usable on AWS Bedrock as an In-Region model in several EU regions Open weights under a Modified MIT licence 196k-token context (per the AWS model card)
Superseded by M2.7, M3 and M3.1 Flash Text input only The MiniMax API ends on 18 November 2026
Deprecated Shutdown: 18 November 2026 (MiniMax API)
MiniMax-M2
October 2025
First generation of the agentic coding model Open source via Hugging Face and GitHub under a Modified MIT licence Available on AWS Bedrock as an In-Region model in several EU regions (until 30 March 2027)
Superseded by M2.5, M2.7, M3 and M3.1 Flash Shut down in the MiniMax API and on Google Agent Platform
Deprecated Shutdown: 30 March 2027 (AWS Bedrock)

Use Cases

Typical applications for this model

Agentic coding & software engineering
Multi-agent workflows
SRE incident response
Complex productivity tasks
Document processing (Word, Excel, PowerPoint)
ML engineering & experimentation
Tool/function calling

Technical Details

API, features and capabilities

API & Availability
Availability Public
Features & Capabilities
Tool Use Function Calling Structured Output Vision Reasoning Mode Code Execution Web Browsing File Upload
Training & Knowledge
Knowledge Cutoff Early 2026
Fine-Tuning Available (Full Fine-tuning, LoRA)
Language Support
Best Quality English, Chinese
Supported Multilingual
Best quality in English and Chinese

Hosting & Compliance

GDPR-compliant hosting options and licensing

GDPR-Compliant Hosting Options
Self-Hosted
EU (your choice)
Deployment on own infrastructure in EU data centres possible (M3.1 Flash, M3 and the M2 series; observe the respective licence)
AWS Bedrock
EU (In-Region selectable, e.g. eu-central-1 Frankfurt, eu-west-1 Ireland, eu-north-1 Stockholm)
MiniMax M2, M2.1 and M2.5 (model IDs minimax.minimax-m2, minimax.minimax-m2.1, minimax.minimax-m2.5) are available on AWS Bedrock only as In-Region inference per the AWS model cards (no Geo or Global cross-region routing) – processing can be pinned to a chosen EU region. M2 and M2.1 are in legacy status with end of life on 30 March 2027; M2.5 is active. M2.7, M3 and M3.1 Flash are not listed on Bedrock.
License & Hosting
License MiniMax Community License (M3.1 Flash, M3); Non-Commercial License (M2.7); Modified MIT (M2, M2.1, M2.5)
Security Filters Configurable
On-Premise

Benchmarks

Performance comparison with standardized tests

DeepSWE v1.1 (M3.1 Flash)
73.7 2026-10
BankerToolBench (M3.1 Flash)
76.2 2026-10
OSWorld 2.0 (M3.1 Flash)
61.7 2026-10
CyberGym (M3.1 Flash)
82.2 2026-10
SWE-Bench Pro (M3)
59 2026-06
Terminal-Bench 2.1 (M3)
66 2026-06
SWE-Pro (M2.7)
56.22 2026
MLE-Bench Lite (Medal Rate, M2.7)
66.6 2026
VIBE-Pro (M2.7)
55.6 2026
Terminal Bench 2 (M2.7)
57 2026
Toolathon (M2.7)
46.3 2026

As an AI consultancy from Rosenheim, we support companies across the DACH region (Germany, Austria, Switzerland) with GDPR-compliant integration of open-weight models such as MiniMax-M3.1-Flash and MiniMax-M3. Through self-hosting in EU data centres, the models can be deployed for agentic workflows and long-context tasks in compliance with data protection regulations. This overview reflects the state as of October 2026.

Current models at a glance

ModelReleasedContextParameters (total / active)InputLicence
MiniMax-M3.1-Flash19 October 20261M tokens~428B / 23BText, image, videoMiniMax Community License
MiniMax-M31 June 20261M tokens~428B / ~23BText, image, videoMiniMax Community License
MiniMax-H3 (video)31 July 2026–33B (H3-Base, dense)Text, image, video, audioMiniMax H3 Community License (EU excluded)

All benchmark figures on this page are vendor-reported by MiniMax unless stated otherwise.

MiniMax-M3.1-Flash

On 19 October 2026, MiniMax released M3.1 Flash – initially via the MiniMax API, the M Plan and MiniMax Code. The weights are now available on Hugging Face (MiniMaxAI/MiniMax-M3.1-Flash). The model builds on M3 and keeps its mixture-of-experts architecture with around 428B parameters, 23B of which are active per token. Per the model card it differs from M3 in three ways:

  • Sparse attention in every layer: the first three full-attention layers of M3 have been replaced with MSA.
  • Quantization-aware training: the routed experts and the attention paths are quantised during training, reducing memory footprint and serving cost.
  • DSpark speculative decoding: the draft head ships with the release and replaces the earlier MTP approach.

MiniMax reports, among others, DeepSWE v1.1 73.7%, BankerToolBench 76.2, OSWorld 2.0 61.7, Agents’ Last Exam 29.6 and CyberGym 82.2. Per MiniMax, output speed exceeds 200 tokens/s in everyday use, with a median above 170 tokens/s. Thinking depth is controlled via effort or reasoning_effort in five levels (low, medium, high, xhigh, max; default max). MiniMax has announced a Pro variant of the M3.1 series but has not released it yet.

MiniMax-M3

MiniMax unveiled MiniMax-M3 on 1 June 2026. The model combines coding and agentic capabilities, a 1-million-token context window and native multimodality (image and video understanding); MiniMax guarantees at least 512k tokens of context in the API.

  • MSA architecture (MiniMax Sparse Attention): per MiniMax, per-token compute at 1M context drops to 1/20 of the predecessor; prefill is more than 9x and decoding more than 15x faster.
  • Native multimodality: per MiniMax, M3 was trained on mixed modalities from the very first step and understands text, images and video in one model – useful for coding with UI screenshots and for document processing.
  • Coding benchmarks (per MiniMax): SWE-Bench Pro 59.0%, Terminal-Bench 2.1 66.0%, MCP Atlas 74.2%.
  • Reasoning: selectable via the thinking parameter (enabled, adaptive, disabled).

The weights are on Hugging Face as MiniMaxAI/MiniMax-M3 (BF16) and MiniMaxAI/MiniMax-M3-MXFP8. Since M3.1 Flash became available, M3 is no longer the first choice on price and efficiency, but it remains available in the API.

MiniMax-H3 (video)

MiniMax-H3 (31 July 2026) is an omni-modal video model that processes text, image, video and audio as a shared context and generates video with native stereo sound. The open H3-Base weights deliver 768p; full 2K quality remains tied to the hosted modules. The MiniMax H3 Community License excludes the EU, the UK, South Korea and the US from the licensed territory, so local operation in the EU is only possible under a separate licence from MiniMax.

Pricing

List prices of the international MiniMax API per 1M tokens (USD):

ModelInputOutputCache read
MiniMax-M3.1-Flash0.150.600.015
MiniMax-M3 (up to 512k input tokens)0.301.200.06
MiniMax-M3 (above 512k input tokens)0.602.400.12

For M3, MiniMax shows the prices as a permanent 50% discount on the original list prices (USD 0.60 / 2.40 up to 512k and USD 1.20 / 4.80 above). The priority tier (service_tier: priority) costs 1.5x the standard price. For M3.1 Flash, MiniMax states a single price with no tiering. Third-party providers and cloud platforms set their own prices.

Prices per provider, as of October 2026.

Licence

M3.1 Flash and M3 are not released under MIT, but under the bespoke MiniMax Community License. Per the LICENSE file in each Hugging Face repository, non-commercial use is free. Any commercial use requires visible attribution (‘Built with MiniMax M3.1-Flash’ or ‘Built with MiniMax M3’). If the resulting products or services exceed US$20 million in annual revenue, prior written authorisation from MiniMax is required; below that threshold a one-time email notice is sufficient. For typical use by DACH mid-market companies the revenue threshold usually does not apply, but the attribution and notice obligations apply regardless of revenue and should be reviewed before rollout.

M2.7 is under a Non-Commercial License modelled on MIT (LICENSE file on Hugging Face, MiniMaxAI/MiniMax-M2.7): private use, research and non-commercial purposes are free, but any commercial use – with no revenue threshold – requires prior written authorisation from MiniMax plus the attribution ‘Built with MiniMax M2.7’.

M2, M2.1 and M2.5 are under a Modified MIT licence per their Hugging Face model cards. Anyone wanting to operate commercially without a separate authorisation from MiniMax should have the relevant licence text reviewed beforehand.

Availability in the EU and GDPR

ModelMiniMax API (storage in the US)AWS BedrockMicrosoft FoundryGoogle Agent PlatformSelf-hosting
MiniMax-M3.1-Flashyesnot listednot listednot listedyes
MiniMax-M3yesnot listedvia Fireworks AI (Global Standard)not listedyes (experimental per MiniMax)
MiniMax-M2.7until 18 Nov 2026, no new usersnot listednot listednot listedyes (Non-Commercial License)
MiniMax-M2.5until 18 Nov 2026, no new usersyes, In-Region incl. Frankfurt, Ireland, Stockholmnot listednot listedyes
MiniMax-M2shut downlegacy until 30 Mar 2027, In-Regionnot listedretired on 21 Oct 2026yes

Who processes the data?

Per its privacy policy (updated 30 March 2026), the international platform (platform.minimax.io, api.minimax.io) is operated by Nanonoble Pte. Ltd., registered in Singapore, as the controller. Per the policy, personal data is stored in a data centre in the United States. MiniMax refers to a cloud provider certified under the EU-US Data Privacy Framework and to standard contractual clauses for transfers to group companies and third parties. The Chinese platform (platform.minimaxi.com) has its own terms.

There is no European Commission adequacy decision for Singapore. Anyone using the API with personal data therefore needs an assessment under Art. 44 et seq. GDPR: a contracting party in Singapore, storage in the US, standard contractual clauses, a transfer impact assessment and a data processing agreement under Art. 28 GDPR. How the cloud provider’s DPF certification affects this transfer chain has to be assessed case by case.

Note: This is a factual summary and does not constitute legal advice. Whether a specific deployment is permissible depends on your role, the categories of data, contractual arrangements and safeguards in the individual case.

Routes for EU deployment

  • Self-hosting: Full data control on your own hardware or on GPU instances in hyperscaler EU regions (e.g. AWS Frankfurt and Ireland, Azure West Europe and Germany West Central, Google Cloud Frankfurt and Belgium) as well as in your own data centre.
  • AWS Bedrock carries MiniMax M2, M2.1 and M2.5 as native models. Per the AWS model cards, only In-Region inference is available (no Geo or Global cross-region routing) – including eu-central-1 (Frankfurt), eu-west-1 (Ireland), eu-south-1 (Milan) and eu-north-1 (Stockholm); eu-west-2 (London) is in the United Kingdom. M2.5 is active, while M2 and M2.1 are in legacy status with end of life on 30 March 2027. M2.7, M3 and M3.1 Flash are not listed on Bedrock.
  • Microsoft Foundry carries MiniMax M3 as a Fireworks on Foundry offering (pay-per-token as Global Standard). Per Microsoft, data is shared with Fireworks AI, data zones are available in the US only, and Fireworks on Foundry is currently excluded from EU Data Boundary commitments.
  • Google Agent Platform: Google retired the MaaS endpoint for MiniMax M2 on 21 October 2026; per Google, all MiniMax MaaS endpoints are deprecated or retired. Self-deployment from Model Garden remains possible.
  • Other inference providers: Together AI, Fireworks AI, OpenRouter. Check regions, retention and terms directly with the provider.

Hardware for self-hosting

With MoE models, all expert weights have to be loaded – 23B active parameters do not mean that memory is only needed for those.

  • MiniMax-M3: per MiniMax’s documentation, self-hosting is still experimental (SGLang development image). The reference is a node with 8x NVIDIA B200 and the MXFP8 weights (around 444 GB); validated configurations also include 4x B300 with MXFP8 and 8x H200 with BF16 weights (around 854 GB).
  • MiniMax-M3.1-Flash: the quantisation-aware trained weights are considerably smaller than M3’s. MiniMax provides recipes for SGLang, vLLM and TokenSpeed but does not name reference hardware in the model card.
  • MiniMax-M2.7: per MiniMax four high-memory GPUs (tensor parallelism 4), with no official minimum.

Migration and breaking changes

  • M2.5 and M2.7 to M3.1 Flash: since 19 October 2026 MiniMax no longer accepts new users for either model family, and both will be shut down on 18 November 2026, after which API calls fail. In the M Plan and the Token Plan, MiniMax has automatically routed requests for M2.5 and M2.7 to M3.1 Flash since 19 October 2026; the existing model names keep working there.
  • Reasoning: M3.1 Flash thinks by default at level max; depth is controlled via effort or reasoning_effort. On the OpenAI-compatible protocol, thinking content is returned separately in reasoning_content. M3 uses the thinking parameter instead (enabled, adaptive, disabled).
  • Cloud platforms: on AWS Bedrock, M2 and M2.1 end on 30 March 2027; the Google MaaS endpoint for M2 has been retired since 21 October 2026.

Model lifecycle

ModelReleasedStatus in the MiniMax APIOther platforms
MiniMax-M3.1-Flash19 October 2026current–
MiniMax-M31 June 2026availableMicrosoft Foundry (Fireworks)
MiniMax-M2.7, M2.7-highspeed18 March 2026no new users since 19 Oct 2026, shutdown on 18 Nov 2026–
MiniMax-M2.5, M2.5-highspeedFebruary 2026no new users since 19 Oct 2026, shutdown on 18 Nov 2026AWS Bedrock active
MiniMax-M2.122 December 2025shut downAWS Bedrock legacy until 30 Mar 2027
MiniMax-M227 October 2025shut downAWS Bedrock legacy until 30 Mar 2027; Google MaaS retired on 21 Oct 2026

The weights of M2, M2.1, M2.5 and M2.7 remain available on Hugging Face (MiniMaxAI) for self-hosting.

Local or Cloud? Calculate the Cost

For agentic coding workloads, a precise cost calculation between API usage and self-hosting often pays off. The local-vs-cloud AI inference calculator from ai-prices.eu lets you compare hardware, electricity and operating costs against cloud API pricing and determine the break-even point for your workload.

Our Recommendation

For new projects: MiniMax-M3.1-Flash, self-hosted in the EU. It is newer and cheaper than M3, comes with smaller weight files and keeps the 1M-token context as well as image and video input. Review the attribution and notice obligations of the MiniMax Community License before rollout.

For existing M3 setups: M3 remains available in the API. Switching to M3.1 Flash pays off mainly for price and hardware requirements; MiniMax itself still classes M3 self-hosting as experimental.

On AWS with a fixed EU region: MiniMax-M2.5 via Bedrock In-Region (e.g. Frankfurt, Ireland, Stockholm) under a Modified MIT licence – an older, text-only generation. M2 and M2.1 only run there until 30 March 2027.

Existing M2.5 and M2.7 users of the MiniMax API should complete their migration before 18 November 2026.

Not recommended for personal or confidential data: the direct MiniMax API without an assessment under Art. 44 et seq. GDPR (contracting party in Singapore, storage in the US) and Fireworks on Foundry while it is excluded from the EU Data Boundary.

We support you with:

  • Infrastructure planning and hardware sizing (especially for 1M-context workloads)
  • Deployment, quantisation and optimisation
  • Integration into existing agent frameworks
  • Compliant usage within CompanyGPT
  • Fine-tuning for specific use cases

MiniMax-M3.1-Flash and M3 combine a 1M-token context, open weights and native multimodality – provided the infrastructure and licence questions are settled.

Get MiniMax-M3 through innFactory

MiniMax-M3 is available via Amazon Bedrock. Which variants run in which regions is described in the EU availability section. innFactory GmbH is a cloud reseller for AWS, Microsoft Azure, Google Cloud and STACKIT: you can source the cloud resources for it directly through us.

  • Billing by invoice or direct debit, no credit card required – all cloud providers on one consolidated invoice.
  • A personal contact at innFactory for setup, EU regions and model selection.
  • innFactory AI Gateway: budgets, SSO and cost control per team and project, as an infrastructure component in your own tenant.
  • CompanyGPT: the GDPR-compliant AI platform for all employees, with free choice of model.

Cost estimation for this model

For up-to-date token pricing, model variants and EU availability, see our sister project ai-prices.eu. It helps you compare and estimate the operational cost of leading AI models for your specific use case.

Compare prices on ai-prices.eu

ai-prices.eu is a project by innFactory AI Consulting GmbH and provides transparent cost estimates for leading AI models.

Consultation for this model?

We help you select and integrate the right AI model for your use case.