MiniMax M2.5 from MiniMax is positioned for agent teams, tool use, and productivity workflows with document editing and coding tasks.
Capabilities, design details, and architectural traits
MiniMax M2.5 is centered on agent workflows and productivity tasks. It is described as a model for complex execution, tool use, and document editing.
| Trait | Documented detail |
|---|---|
| Agent teams | It is presented for complex agent harnesses and coordinated agent work. |
| Tool use | It supports tool interaction and dynamic tool search during execution. |
| Office editing | It is documented for multi-turn editing in Excel, PowerPoint, and Word. |
| Coding and engineering | It is highlighted for software delivery, bug hunting, security analysis, and ML tasks. |
The model is framed for tasks that need repeated execution and coordination, not only single-turn chat. Its distinguishing pattern is the combination of tool use, agent coordination, and structured office work.
Its public description emphasizes software engineering and document editing. That makes it a workflow-first model rather than a general-purpose wording of capability.
Independent evaluations · Artificial Analysis
Evaluate specifications, pricing, and independent benchmark indices
| Model Details | |||
|---|---|---|---|
| General Info | |||
| Provider | MiniMax | Anthropic | Anthropic |
| Release Date | February 12, 2026 | July 24, 2026 | June 9, 2026 |
| Knowledge Cutoff | - | May 2026 | - |
| Context & Limits | |||
| Context Window | 205K | 1M Best Context Window | 1M Best Context Window |
| Pricing (per 1M tokens) | |||
| Input Pricing | $0.30 Best Input Pricing | $5 | $10 |
| Output Pricing | $1.20 Best Output Pricing | $25 | $50 |
| Modalities | |||
| Inputs | text | textimage | textimagefile |
| Outputs | text | text | text |
| Benchmarks (0-100) | |||
| Intelligence Index | 34.5 | 63.1 Best Intelligence Index | 62.1 |
| Coding Index | - | 78.0 Best Coding Index | 76.5 |
| Agentic Index | - | 59.2 Best Agentic Index | 56.6 |
Graduate-level reasoning and expert Q&A evaluation.
Extremely difficult logical reasoning and knowledge.
Logical reasoning over long context windows.
Independent evaluation data provided by Artificial Analysis. To view the latest benchmarks and full details, visit their official site.
