LLM Reference

MiniMax M3

Released
2026-06-01
Last refreshed
2026-06-30
Status
Researched 66d ago
Open weightsCommercial use: conditionalMultimodalCodingRAGAgentsLong contextVisionJSON / Tool useHighlight

MiniMax M3 is worth evaluating for coding, rag, and agents when its provider route and context window match the workload.

Use it for

  • Teams evaluating coding, rag, and agents
  • Workloads that can use a 1m context window
  • Buyers comparing 2 tracked provider routes

Do not use it for

  • Workloads where another current model has stronger sourced task evidence
Specifications
Released
2026-06-01
Context
1m
Max output
131,072
Architecture
Decoder Only
Specialization
general
Openness
Open weights
License
MiniMax Community LicenseCommercial use: conditional
Weights
Available
Code
Available·MITOSI-approved
Created by
MiniMaxMiniMax

Developing AI for gaming and entertainment.

Minhang, Shanghai, China
Founded 2021
Website
Pricing
Output / 1M
$1.20
Input / 1M
$0.300

Cheapest of 2 routes · MiniMax · cache read $0.060

Providers(2)
MiniMaxOpenRouter
View 2 provider routes
Links
Website HuggingFace

About

MiniMax M3 is MiniMax's current API flagship (released June 1, 2026) with MiniMax Sparse Attention for economical 1M-token context, native multimodality, and agentic coding. Benchmark rows in LLMReference separate vendor-reported SWE-bench Pro, Terminal-Bench 2.1, MCP-Atlas, and BrowseComp scores (source: minimax.io) from third-party rows — inspect evaluator, variant, and source on each benchmark cell before comparing to leaderboard claims.

Top use-case fit: coding, agents, and build tasks

Coding

Q/$ C

2 relevant benchmarks in the decision map.

RAG

Included by capability and metadata signals in the decision map.

Agents

Q/$ B

1 relevant benchmark in the decision map.

Provider price ladder

Compare all 2

Compare API pricing across 2 providers for input and output tokens, batch, and cached reads when available.

ProviderInput / 1MOutput / 1MCacheRoute
MiniMax$0.300$1.20read $0.060
Serverless
OpenRouter$0.300$1.20-
Serverless

Capabilities

VisionMultimodalReasoningJSON / Tool useStructured OutputsCode ExecutionPrompt Caching

Benchmark peer barsfor Coding

SWE-bench ProRank 14 of 46
Claude Fable 5
80.3
Claude Opus 5
79.2
Claude Opus 4.8
69.2
Qwen3.8-Max
67.7
MiniMax M3current
59.0
SWE-bench VerifiedRank 14 of 81
Claude Fable 5
96.0
Claude Opus 5
96.0
Claude Mythos Preview
93.9
Claude Opus 4.8
88.6
MiniMax M3current
80.5

Benchmark scores(9)

Scores are benchmark-specific and are direction-aware: the same numeric gap can mean very different outcomes across suites. Use the leaderboard context and this model's provider route to decide whether the winning margin is meaningful for your workload.
BenchmarkScoreVersionEvaluationSource
MMMU Pro78.1LLM-Stats aggregatorObserved 2026-06-07—Source
SWE-bench Pro59.0MiniMax-reportedObserved 2026-06-01—Source
Terminal-Bench 2.166.0MiniMax-reportedObserved 2026-06-01—Source
MCP-Atlas74.2MiniMax-reported public setObserved 2026-06-01—Source
BrowseComp83.5MiniMax-reportedObserved 2026-06-01—Source
Google-Proof Q&A92.9MiniMax M3 GPQA 92 (accuracy%)Observed 2026-06-07—Source
SWE-bench Verified80.5Rank 8 of 49 (resolved%)Observed 2026-06-07—Source
Terminal-Bench 2.066.0Terminal-Bench 2.0 (accuracy%)Observed 2026-06-07—Source
GeneBench-Pro0.9xhighObserved 2026-06-30—Source

Migration checks

No linked migration route is available for this model yet.

Compare MiniMax M3 with other models