MiniMax M2.5 by MiniMax demonstrates top-tier general intelligence, competitive pricing. View detailed benchmark data including scores across coding, math, reasoning, speed, and cost metrics.
Specifications last checked against the provider on . Benchmark scores carry their own source, linked beside each number.
General Benchmarks
Human preference ELO from blind head-to-head votes
LMSYS / HuggingFaceCoding Benchmarks
Human preference ELO for building things — websites, UI, games, charts, SVG
Design Arena (via OpenRouter)Human preference ELO for building a working web page from a brief
Design Arena (via OpenRouter)Reasoning Benchmarks
Multi-turn service agent making tool calls under strict policy constraints
OpenRouter (measured)Speed Benchmarks
Cost Benchmarks
Cost per 1M cached input tokens — the price that actually applies to a long agent conversation
OpenRouter APIContext Benchmarks
Available from 45 providers
The same model costs different amounts depending on who serves it. You can bring your own key for 8 of these — connect it here.
| Provider | Input / 1M | Output / 1M | Cached in | Context | Your key |
|---|---|---|---|---|---|
| Alibaba Coding Plan | $0 | $0 | $0 | 197K | — |
| Alibaba Coding Plan (China) | $0 | $0 | $0 | 197K | — |
| Alibaba Token Plan | $0 | $0 | $0 | 197K | — |
| Alibaba Token Plan (China) | $0 | $0 | $0 | 197K | — |
| MiniMax Token Plan (minimaxi.com) | $0 | $0 | $0 | 205K | — |
| MiniMax Token Plan (minimax.io) | $0 | $0 | $0 | 205K | — |
| SCNet Token Plan | $0 | $0 | $0 | 205K | — |
| Tencent Coding Plan (China) | $0 | $0 | $0 | 205K | — |
| Deep Infra | $0.15 | $1.15 | $0.03 | 197K | Supported |
| GreenPT | $0.1938 | $1.129 | $0.0627 | 205K | — |
| DInference | $0.22 | $0.88 | — | 200K | — |
| OpenRouter | $0.27 | $1.08 | $0.027 | 205K | Supported |
| Venice AI | $0.27 | $0.95 | $0.03 | 198K | — |
| D.Run (China) | $0.29 | $1.16 | — | 205K | — |
| Cortecs | $0.296 | $1.186 | $0.075 | 196K | — |
| 302.AI | $0.3 | $1.2 | — | 205K | — |
| Alibaba (China) | $0.3 | $1.2 | — | 205K | — |
| Amazon Bedrock | $0.3 | $1.2 | — | 197K | Supported |
| Baseten | $0.3 | $1.2 | — | 204K | Supported |
| CloudFerro Sherlock | $0.3 | $1.2 | — | 196K | — |
| DevPass (LLM Gateway) | $0.3 | $1.2 | $0.03 | 229K | — |
| DigitalOcean | $0.3 | $1.2 | $0.06 | 66K | — |
| Eden AI | $0.3 | $1.2 | $0.03 | 205K | — |
| Friendli | $0.3 | $1.2 | $0.06 | 197K | — |
| FrogBot | $0.3 | $1.2 | $0.03 | 192K | — |
| HPC-AI | $0.3 | $1.2 | $0.03 | 196K | — |
| Hugging Face | $0.3 | $1.2 | $0.03 | 205K | Supported |
| Kilo Gateway | $0.3 | $1.2 | $0.03 | 200K | — |
| LLM Gateway | $0.3 | $1.2 | $0.03 | 205K | — |
| Meganova | $0.3 | $1.2 | — | 205K | — |
| Merge Gateway | $0.3 | $1.2 | $0.03 | 205K | — |
| MiniMax (minimaxi.com) | $0.3 | $1.2 | $0.03 | 205K | — |
| MiniMax (minimax.io) | $0.3 | $1.2 | $0.03 | 205K | Supported |
| NanoGPT | $0.3 | $1.2 | $0.15 | 205K | — |
| NovitaAI | $0.3 | $1.2 | $0.03 | 205K | — |
| Ofox | $0.3 | $1.2 | $0.03 | 205K | — |
| OpenCode Go | $0.3 | $1.2 | $0.03 | 205K | — |
| OpenCode Zen | $0.3 | $1.2 | $0.06 | 205K | — |
| OrcaRouter | $0.3 | $1.2 | $0.03 | 205K | — |
| TensorX | $0.3 | $1.2 | $0.075 | 197K | — |
| Together AI | $0.3 | $1.2 | $0.06 | 205K | Supported |
| TokenGo | $0.3 | $1.2 | $0.03 | 205K | — |
| Vercel AI Gateway | $0.3 | $1.2 | $0.03 | 205K | Supported |
| ZenMux | $0.3 | $1.2 | $0.03 | 205K | — |
| Ollama Cloud | — | — | — | 205K | — |
MiniMax M2.5 — Benchmark Scores Overview
Scores normalized to percentage scale for visual comparison. ELO scores mapped to 0-100 range (1100-1500).
Compare MiniMax M2.5 With
MiniMax M2.5 — Frequently Asked Questions
How intelligent is MiniMax M2.5?
MiniMax M2.5 scores 1443 on the Chatbot Arena ELO rating, making it a high-performing AI model. This score is based on blind head-to-head human preference voting.
How much does MiniMax M2.5 cost?
MiniMax M2.5 costs $0.27 per 1M input tokens and $1.1 per 1M output tokens. This makes it one of the more affordable models.
How fast is MiniMax M2.5?
MiniMax M2.5 generates output at 59 tokens per second, which is slower, prioritizing quality over speed compared to other models. The time to first token is 2330 ms.
What is the context window of MiniMax M2.5?
MiniMax M2.5 has a context window of 205K tokens. This determines how much text, conversation history, and code the model can process in a single request.
Who created MiniMax M2.5?
MiniMax M2.5 was created by MiniMax. It is classified as a flagship model in our catalogue.
Is MiniMax M2.5 open source?
No, MiniMax M2.5 is a proprietary model. It is available through MiniMax's API and compatible providers.