LLM Reference

MiniMax M2 vs MiniMax M3

MiniMax M2 (2025) and MiniMax M3 (2026) are frontier reasoning models from MiniMax. MiniMax M2 ships a 197k-token context window, while MiniMax M3 ships a 1m-token context window. On pricing, MiniMax M2 costs $0.26/1M input tokens; MiniMax M3 ranges from $0.30 to $0.60/1M input tokens by tier. This comparison covers specs, pricing, API access, capabilities, benchmarks, input and output token costs, and production fit for coding and agent workloads.

MiniMax M3 fits 5x more tokens; pick it for long-context work and MiniMax M2 for tighter calls.

Decision scorecard

Local evidence first
SignalMiniMax M2MiniMax M3
Best forprovider-routed productionreasoning-heavy apps, multimodal apps, and tool-calling agents
Decision fitRAG, Long context, and ClassificationCoding, RAG, and Agents
Context window197k1m
Cheapest output$1/1M tokens$1.20/1M tokens
Provider routes5 tracked2 tracked
Shared benchmarks0 shared0 shared

Decision tradeoffs

Choose MiniMax M2 when...
  • MiniMax M2 has the lower cheapest tracked output price at $1/1M tokens.
  • MiniMax M2 has broader tracked provider coverage for fallback and procurement flexibility.
  • Local decision data tags MiniMax M2 for RAG, Long context, and Classification.
Choose MiniMax M3 when...
  • MiniMax M3 has the larger context window for long prompts, retrieval packs, or transcript analysis.
  • MiniMax M3 uniquely exposes Vision, Multimodal, and Reasoning in local model data.
  • Local decision data tags MiniMax M3 for Coding, RAG, and Agents.

Monthly cost at traffic

Estimate token spend from the cheapest tracked input and output route or tier on this page.

Lower estimate MiniMax M2

MiniMax M2

$454

Cheapest tracked route/tier: OpenRouter

MiniMax M3

$540

Cheapest tracked route/tier: MiniMax <=512K input tokens (standard)

Estimated monthly gap: $86.00. Batch, cache, alternate speed tiers, and negotiated pricing are excluded from this local estimate.

Switch friction

MiniMax M2 -> MiniMax M3
  • Provider overlap exists on OpenRouter; start route-level A/B tests there.
  • MiniMax M3 is $0.20/1M tokens higher on cheapest tracked output pricing, so quality gains need to justify the spend.
  • MiniMax M3 adds Vision, Multimodal, and Reasoning in local capability data.
MiniMax M3 -> MiniMax M2
  • Provider overlap exists on OpenRouter; start route-level A/B tests there.
  • MiniMax M2 is $0.20/1M tokens lower on cheapest tracked output pricing before cache, batch, or negotiated discounts.
  • Check replacement coverage for Vision, Multimodal, and Reasoning before moving production traffic.

Specs

Specification
Released2025-10-012026-06-01
Context window197k1m
Parameters230B (10B active)
ArchitectureDecoder OnlyDecoder Only
LicenseMITOSI-approvedMiniMax Community License
OpennessOpen sourceOpen weights
WeightsUnknownAvailable
CodeUnknownAvailable·MITOSI-approved
Commercial useCommercial use: permittedCommercial use: conditional
Knowledge cutoff--

Pricing and availability

Pricing attributeMiniMax M2MiniMax M3
Input price$0.26/1M tokens
<=512K input tokens (standard)
$0.30/1M tokens
Permanent standard-tier price for <=512K input tokens; cache_read stores this tier's $0.06/M cache-read price.
>512K input tokens
$0.60/1M tokens
Long-context tier for inputs above 512K; cache-read price is $0.12/M per MiniMax pricing docs.
Output price$1/1M tokens
<=512K input tokens (standard)
$1.20/1M tokens
Permanent standard-tier price for <=512K input tokens; cache_read stores this tier's $0.06/M cache-read price.
>512K input tokens
$2.40/1M tokens
Long-context tier for inputs above 512K; cache-read price is $0.12/M per MiniMax pricing docs.
Providers

Capabilities

CapabilityMiniMax M2MiniMax M3
VisionNoYes
MultimodalNoYes
ReasoningNoYes
Function callingNoYes
Tool useNoYes
Structured outputsYesYes
Code executionNoYes
IDE integrationNoNo
Computer useNoNo
Parallel agentsNoNo

Benchmarks

No shared benchmark scores are currently available for this pair.

Deep dive

The capability footprint differs most on vision: MiniMax M3, multimodal input: MiniMax M3, reasoning mode: MiniMax M3, function calling: MiniMax M3, tool use: MiniMax M3, and code execution: MiniMax M3. Both models share structured outputs, so the practical split is not just feature count. Use those differences to decide whether the page is about raw model quality, agentic coding support, multimodal ingestion, or predictable structured API behavior.

For cost, MiniMax M2 lists $0.26/1M input and $1/1M output tokens on the cheapest tracked provider, while MiniMax M3 lists tiered pricing: <=512K input tokens (standard) is $0.30/1M input and $1.20/1M output; >512K input tokens is $0.60/1M input and $2.40/1M output. A 70/30 input-output blend puts MiniMax M2 lower by about $0.09 per million blended tokens. For tiered rows, this cheapest-track view can understate interactive or fast-lane spend, so compare the tier you will actually use. Availability is 5 providers versus 2, so concentration risk also matters.

Choose MiniMax M2 when provider fit, lower input-token cost, and broader provider choice are central to the workload. Choose MiniMax M3 when coding workflow support and larger context windows are more important. For production, rerun your own prompts through the exact provider, region, and tool stack you plan to ship.

FAQ

Which has a larger context window, MiniMax M2 or MiniMax M3?

MiniMax M3 supports 1m tokens, while MiniMax M2 supports 197k tokens. That gap matters most for long documents, large codebases, retrieval-heavy agents, and conversations where earlier context must remain visible.

Which is cheaper, MiniMax M2 or MiniMax M3?

MiniMax M2 lists $0.26/1M input and $1/1M output tokens on the cheapest tracked provider. MiniMax M3 lists tiered pricing: <=512K input tokens (standard) is $0.30/1M input and $1.20/1M output; >512K input tokens is $0.60/1M input and $2.40/1M output. Compare the tier you will actually use; cheap async pricing can overstate savings for interactive workflows. Provider discounts or batch pricing can still change the final bill.

Is MiniMax M2 or MiniMax M3 open source?

MiniMax M2 is listed under MIT. MiniMax M3 is listed under MiniMax Community License. License labels affect whether you can self-host, redistribute weights, or rely only on hosted APIs, so confirm the upstream license before deployment.

Which is better for vision, MiniMax M2 or MiniMax M3?

MiniMax M3 has the clearer documented vision signal in this comparison. If vision is mission-critical, validate it against the provider endpoint because model-level support and API-level exposure can differ. Use this as a quick comparison signal, then confirm the provider-specific limits before committing to production.

Which is better for multimodal input, MiniMax M2 or MiniMax M3?

MiniMax M3 has the clearer documented multimodal input signal in this comparison. If multimodal input is mission-critical, validate it against the provider endpoint because model-level support and API-level exposure can differ.

Where can I run MiniMax M2 and MiniMax M3?

MiniMax M2 is available on Fireworks AI, AWS Bedrock, OpenRouter, Vercel AI Gateway, and Novita AI. MiniMax M3 is available on MiniMax and OpenRouter. Provider coverage can affect latency, region availability, compliance posture, and fallback options.

Continue comparing

Last reviewed: 2026-06-30. Data sourced from public model cards and provider documentation.