MiMo-V2.5-Pro
Xiaomi MiMo-V2.5-Pro is an open-source Mixture-of-Experts model from Xiaomi for agentic coding, long-horizon tasks, and a 1M-token context window.
Model Overview
Capabilities, design details, and architectural traits
MiMo-V2.5-Pro - 1M-token agentic long-horizon model
MiMo-V2.5-Pro is built for long, tool-heavy work. It is designed for agentic tasks, complex software engineering, and long-horizon trajectories that stay coherent across very large context windows.
| Differentiator | Documented trait |
|---|---|
| 1M-token context | Supports up to 1M tokens of context. |
| Mixture-of-Experts design | Uses a Mixture-of-Experts (MoE) model. |
| Hybrid attention + MTP | Uses hybrid attention and Multi-Token Prediction from the MiMo-V2-Flash design. |
| Long tool-call trajectories | Can sustain complex trajectories spanning thousands of tool calls. |
Why it stands out
The model pairs very long context with agent-focused behavior. That makes it different from a standard general chat model, because the documented emphasis is on extended task execution and strong instruction following across long runs.
It is also open-sourced and presented as the most capable MiMo model in the series. The model card describes it as suited to the most demanding agentic, software engineering, and long-horizon tasks.
Benchmark Performance
Independent evaluations · Artificial Analysis
Accuracy & Capability Details
Compare Models Side-by-Side
Evaluate specifications, pricing, and independent benchmark indices
| Model Details | |||
|---|---|---|---|
| General Info | |||
| Provider | Xiaomi | Anthropic | Anthropic |
| Release Date | April 22, 2026 | September 22, 2026 | September 28, 2026 |
| Knowledge Cutoff | - | - | Jun 2026 |
| Context & Limits | |||
| Context Window | 1.0M Best Context Window | 1M | 1M |
| Pricing (per 1M tokens) | |||
| Input Pricing | $0.43 Best Input Pricing | $4 | $2 |
| Output Pricing | $0.87 Best Output Pricing | $20 | $10 |
| Modalities | |||
| Inputs | text | textimagefile | textimagefile |
| Outputs | text | text | text |
| Benchmarks (0-100) | |||
| Intelligence Index | 26.0 | 57.6 Best Intelligence Index | 56.0 |
| Coding Index | 60.2 | - | - |
| Agentic Index | 21.3 | - | - |
Humanity's Last Exam
Extremely difficult logical reasoning and knowledge.
Long Context Reasoning
Logical reasoning over long context windows.
SciCode Benchmark
Scientific coding and mathematical modeling.
Independent evaluation data provided by Artificial Analysis. To view the latest benchmarks and full details, visit their official site.
Explore more from Xiaomi
Other models by Xiaomi
MiMo-V2.6-Pro
MiMo-V2.6-Flash
MiMo-V2.5
MiMo-V2-Flash
Top AI Models
Leading alternatives by intelligence score
