MiniMax-M2.5
As of Oct 4, 2026, the cheapest MiniMax-M2.5 API price is $0.1938 per 1M input tokens (GreenPT) and $0.88 per 1M output tokens (DInference). MiniMax-M2.5 is offered by 41 providers in the minimax family, open weights, reasoning enabled. MiniMax (minimax.io)'s own rate is $1.2 per 1M output tokens. It is also included in 6 flat-rate subscription plans, Qwen (Alibaba Cloud) Token Plan Lite from $8/mo.
API pricing by provider
Metered per-token prices. Cheapest input and output are highlighted; † marks a context-tier surcharge. How prices are ranked.
| DInferenceopen | $0.22 | Cheapest$0.88 | — | — | 200k | — |
| GreenPTopen | Cheapest$0.1938 | $1.129 | $0.0627 | — | 205k | — |
| Venice AIopen | $0.27 | $0.95 | $0.03 | — | 198k | — |
| OpenRouteropen | $0.27 | $1.08 | $0.027 | — | 205k | 21d agoout −10% |
| D.Run (China) | $0.29 | $1.16 | — | — | 205k | — |
 302.AIopen | $0.3 | $1.2 | — | — | 205k | — |
| Alibaba (China)open | $0.3 | $1.2 | — | — | 205k | — |
| Amazon Bedrockopen | $0.3 | $1.2 | — | — | 197k | — |
| CloudFerro Sherlockopen | $0.3 | $1.2 | — | — | 196k | — |
| Eden AIopen | $0.3 | $1.2 | $0.03 | — | 205k | — |
| Friendliopen | $0.3 | $1.2 | $0.06 | — | 197k | — |
 FrogBot | $0.3 | $1.2 | $0.03 | — | 192k | — |
| HPC-AIopen | $0.3 | $1.2 | $0.03 | — | 196k | — |
| Hugging Faceopen | $0.3 | $1.2 | $0.03 | — | 205k | — |
| Kilo Gatewayopen | $0.3 | $1.2 | $0.03 | — | 200k | — |
| DevPass (LLM Gateway)open | $0.3 | $1.2 | $0.03 | $0.375 | 229k | — |
| LLM Gatewayopen | $0.3 | $1.2 | $0.03 | — | 205k | — |
| LLM Gatewayopen | $0.3 | $1.2 | $0.03 | — | 205k | — |
| Meganovaopen | $0.3 | $1.2 | — | — | 205k | — |
| Merge Gatewayopen | $0.3 | $1.2 | $0.03 | $0.375 | 205k | — |
| MiniMax (minimax.io)open | $0.3 | $1.2 | $0.03 | $0.375 | 205k | — |
| MiniMax (minimax.cn)open | $0.3 | $1.2 | $0.03 | $0.375 | 205k | — |
| NanoGPTopen | $0.3 | $1.2 | $0.15 | — | 205k | — |
| NovitaAI | $0.3 | $1.2 | $0.03 | — | 205k | — |
| Ofoxopen | $0.3 | $1.2 | $0.03 | $0.375 | 205k | — |
| OpenCode Zenopen | $0.3 | $1.2 | $0.06 | — | 205k | — |
| OrcaRouteropen | $0.3 | $1.2 | $0.03 | $0.375 | 205k | — |
| TensorXopen | $0.3 | $1.2 | $0.075 | $0.375 | 197k | — |
| TokenGoopen | $0.3 | $1.2 | $0.03 | — | 205k | — |
| Vercel AI Gatewayopen | $0.3 | $1.2 | $0.03 | $0.375 | 205k | — |
| ZenMux | $0.3 | $1.2 | $0.03 | $0.375 | 205k | — |
| Cortecsopen | $0.348 | $1.39 | $0.075 | — | 196k | todayout +17.2% |
Subscription plans (not per-token)
Billed per month, not per token — never counted as the cheapest offer.
| Alibaba Coding Planplan | free | free | 197k |
| Alibaba Coding Plan (China)plan | free | free | 197k |
| Alibaba Token Planplan | free | free | 197k |
| Alibaba Token Plan (China)plan | free | free | 197k |
| MiniMax Token Plan (minimax.cn)plan | free | free | 205k |
| MiniMax Token Plan (minimax.io)plan | free | free | 205k |
| SCNet Token Planplan | free | free | 205k |
| Tencent Coding Plan (China)plan | free | free | 205k |
MiniMax-M2.5 pricing FAQ
What is the cheapest MiniMax-M2.5 API?
As of Oct 4, 2026, DInference has the lowest MiniMax-M2.5 output price at $0.88 per 1M tokens, and GreenPT has the lowest input price at $0.1938 per 1M tokens. Prices are metered per-token rates in USD; free tiers and subscription plans are excluded.
How much does MiniMax-M2.5 cost on MiniMax (minimax.io)?
MiniMax (minimax.io) charges $0.3 per 1M input tokens and $1.2 per 1M output tokens for MiniMax-M2.5.
How many providers offer MiniMax-M2.5?
41 providers list MiniMax-M2.5 on Sovyron; 32 of them sell it at a metered per-token price.
Is MiniMax-M2.5 free?
No provider in the Sovyron catalog lists a free tier for MiniMax-M2.5.
Is MiniMax-M2.5 included in a subscription plan?
Yes. 6 flat-rate plan(s) include it; the cheapest is Qwen (Alibaba Cloud) Token Plan Lite at $8/month.
What is the context window of MiniMax-M2.5?
MiniMax-M2.5 supports a 229k-token context window and up to 196k output tokens.
Token cost calculator
What a month costs at one provider's published rates. Enter the tokens you expect to send and receive; cached tokens use the cached-input price where the provider lists one, and count as ordinary input where it does not.
Estimated
—