Skip to content
LLMs
MiniMax logo

MiniMax: MiniMax M2.5

by MiniMax

MiniMax: MiniMax M2.5 is a large language model from MiniMax. It costs $0.270 per million input tokens and $1.08 per million output tokens. Its context window is 205K tokens.

ReasoningTool callingStructured outputPrompt cachingOpen weights
Input / 1M tokens
$0.270
Output / 1M tokens
$1.08
Cached input / 1M
$0.027

On repeated prefixes

Context window
205K

tokens

Who serves it cheapest

7 hosts serve MiniMax M2.5. Same weights, same API - the price difference is pure margin and routing.

Providers serving MiniMax: MiniMax M2.5, cheapest first
ProviderInput / 1MOutput / 1MContextThroughputUptime 24h
VeniceCheapest$0.270$0.950198K-75.7%
StreamLake$0.270$1.08200K-99.7%
AtlasCloudfp8$0.295$1.20197K-99.5%
DigitalOcean$0.300$1.2066K-100.0%
Friendli$0.300$1.20197K-100.0%
Novitafp8$0.300$1.20205K-99.9%
Minimaxfp8$0.300$1.20205K-100.0%

The spread between Venice and Minimax is 1.2× for identical weights. Quantization and context limits differ, so check both columns before switching.

About MiniMax M2.5

MiniMax-M2.5 is a SOTA large language model designed for real-world productivity. Trained in a diverse range of complex real-world digital working environments, M2.5 builds upon the coding expertise of M2.1...

Specifications

MiniMax: MiniMax M2.5 specifications
Model IDminimax/minimax-m2.5
ProviderMiniMax
Context window205K tokens
Max output128K tokens
Input modalitiestext
Output modalitiestext
Knowledge cutoff-
Open weightsYes - MiniMaxAI/MiniMax-M2.5
ReleasedFebruary 12, 2026

Cheaper alternatives

Models that cost less than MiniMax M2.5 while keeping at least half its context window and every input modality it supports.

Frequently asked

How much does MiniMax: MiniMax M2.5 cost?

$0.270 per million input tokens and $1.08 per million output tokens. Cached input reads cost $0.027 per million tokens.

What is the context window of MiniMax: MiniMax M2.5?

205K tokens, with up to 128K tokens of output per request.

Which provider serves MiniMax: MiniMax M2.5 cheapest?

Venice at $0.270 per million input tokens - 1.2× cheaper than Minimax, the most expensive of the 7 hosts serving it.

Confirm against the source: MiniMax official pricing.