Claude Fable 5$22.000/MClaude Opus 4.8$11.000/MClaude Opus 4.7$11.000/MClaude Opus 4.6$11.000/MClaude Opus 4.5$33.000/MClaude Sonnet 3.7$6.600/MClaude Opus 3$33.000/MClaude 2.1$12.800/MClaude 2$12.800/MGPT-5.5$12.500/MGPT-5.2$5.425/MGPT-5.2-Codex$5.425/MGPT-5$3.875/MGPT-4.5$97.500/MGPT-4 Turbo Preview$16.000/MGPT-4$39.000/MGPT-4-32k$78.000/Mo3$19.000/Mo3-mini$2.090/Mo4-mini$2.090/Mo1$28.500/Mo1-mini$5.700/Mo1-preview$28.500/MGemini 3.5 Pro$5.000/MGemini 3.1 Pro$5.000/MGemini 3 Pro$5.000/MGemini 2.5 Pro$3.875/MGemini 1.5 Pro$2.375/MGemini 1.0 Ultra$12.000/MGemini 1.0 Pro$0.800/MClaude Fable 5$22.000/MClaude Opus 4.8$11.000/MClaude Opus 4.7$11.000/MClaude Opus 4.6$11.000/MClaude Opus 4.5$33.000/MClaude Sonnet 3.7$6.600/MClaude Opus 3$33.000/MClaude 2.1$12.800/MClaude 2$12.800/MGPT-5.5$12.500/MGPT-5.2$5.425/MGPT-5.2-Codex$5.425/MGPT-5$3.875/MGPT-4.5$97.500/MGPT-4 Turbo Preview$16.000/MGPT-4$39.000/MGPT-4-32k$78.000/Mo3$19.000/Mo3-mini$2.090/Mo4-mini$2.090/Mo1$28.500/Mo1-mini$5.700/Mo1-preview$28.500/MGemini 3.5 Pro$5.000/MGemini 3.1 Pro$5.000/MGemini 3 Pro$5.000/MGemini 2.5 Pro$3.875/MGemini 1.5 Pro$2.375/MGemini 1.0 Ultra$12.000/MGemini 1.0 Pro$0.800/M
BETA
Subscribe
← Back to Index·All MiniMax models
MiniMax
MiniMaxFrontierLIVE INDEX

MiniMax M2.5

text->text

MiniMax's flagship hybrid lightning-attention model. 1M-token context with strong Chinese coverage and competitive frontier pricing.

MiniMax M2.5 is a frontier AI model from MiniMax. It costs $0.150 per million input tokens and $0.900 per million output tokens (blended $0.375/M), with a 205K-token context window.

Profile inherited from upstream MiniMax-Text-01 (M2 generation) ↗ — this is a hosted variant of the same open-weights model.

Released Jan 2025Modalities textOfficial model page ↗Pricing source ↗Compare with another model →Estimate monthly cost →
INPUT
$0.150/M
per million input tokens
OUTPUT
$0.900/M
per million output tokens
BLENDED 70/30
$0.375/M
default reference rate · how it's calculated →
CONTEXT
205K
204,800 tokens
What it's good at
  • 1M context
  • Lightning attention (linear-time)
  • Chinese-tuned
  • Competitive frontier quality
Typical use cases
  • Long Chinese document RAG
  • Cheap long-context inference
  • Multilingual chat
Benchmarks
vs. best public score
Scores inherited from MiniMax-Text-01 (M2 generation) — this is a hosted variant of the same open-weights model, so the underlying benchmark scores are identical.
MMLU82%
Multitask academic knowledge across 57 subjects.
MATH78%
High-school competition math problems.
HumanEval80%
Python function synthesis from docstrings.
LMArena Elo1232 Elo
Crowd-sourced head-to-head preference Elo rating.
Hand-curated from each provider's published reports and public leaderboards. Methodology varies across sources — treat as directional rather than authoritative.
More from MiniMax
See all 5 →
MiniMax-Text-01
Frontier · 1M ctx
in $0.700/Mout $2.800/M
ABAB 6.5s
Balanced · 8K ctx
in $1.000/Mout $5.000/M
ABAB 6.5t
Turbo · 8K ctx
in $0.450/Mout $0.450/M
ABAB 5.5s
Efficient · 16K ctx
in $0.150/Mout $0.750/M
ABAB 5.5c
Compact · 16K ctx
in $0.050/Mout $0.050/M
Frequently asked questions

How much does MiniMax M2.5 cost?

MiniMax M2.5 costs $0.150 per million input tokens and $0.900 per million output tokens, for a blended reference rate of $0.375 per million tokens.

What is MiniMax M2.5's context window?

MiniMax M2.5 supports up to 205K tokens of context (204,800 tokens).

What is MiniMax M2.5 best for?

MiniMax M2.5 is well suited to 1M context, Lightning attention (linear-time) and Chinese-tuned.

Who makes MiniMax M2.5?

MiniMax M2.5 is developed and served by MiniMax. It was released in Jan 2025.