Xiaomi MiMo-V2.5
Model overview
Xiaomi MiMo-V2.5
Xiaomi MiMo-V2.5 omni-modal model with parameter labeling on the first-party Hugging Face organization page.
general LLMreasoningmultimodalvision-languagelong contextchatdocument understandinglocal inferenceevaluation
Representative version
Xiaomi MiMo-V2.5
| Company | Xiaomi |
|---|---|
| Release | 2026-04-23 |
| Release type | Model release |
| Parameters | 310B / 15B active |
| Architecture | MoE or sparse architecture; exact architecture not separately normalized. |
| Context | 1,000,000 |
| Input modalities | text / image / video / audio |
| Output modalities | text / code / tool_call |
| Tool use / function calling | Not disclosed |
| Structured output / reasoning | Confirmed: Reasoning |
| Vision / audio / video | Confirmed: Vision, Audio, Video |
| Open weights / License | Weights available / MIT License sourcexiaomi-mimo-v2-5-phase-3-5 · 2026-07-15 |
| Vendor API | Available Xiaomi MiMo-V2.5 Hugging Face model cardxiaomi-mimo-v2-5-phase-3-5 · 2026-04-23 |
| OpenRouter | OpenRouter listed |
| Pricing | OpenRouter · 2026-07-15: $0.14 / $0.28 input / output, per 1M tokens |
| Benchmark | LMArena 4 rows |
| Public signals | 3 runtime / access records / 1 activity snapshot |
| Use cases | assistant chat / document understanding / local inference / research evaluation |
| Claim evidence | Release / model specs / weightsXiaomi MiMo-V2.5 Hugging Face model cardXiaomi MiMo-V2.5 · 2026-04-23 |
OpenRouter
Pricing / context / availability
| OpenRouter | OpenRouter listed |
|---|---|
| OpenRouter listing | xiaomi/mimo-v2.5 |
| Mapped version | Xiaomi MiMo-V2.5 xiaomi-mimo-v2-5-phase-3-5 |
| Pricing | OpenRouter · 2026-07-15: $0.14 / $0.28 input / output, per 1M tokens |
| Context | 1048576 |
| Speed/Latency | No reliable public source |
| Snapshot time | 2026-07-15T20:11:41.551Z |
| Source | View source |
Benchmark
Ranking evidence
| Source | Task | Rank | Score | Sample | Time | Link |
|---|---|---|---|---|---|---|
| LMArena | WebDev | 18 | 1,442.8 | 2,198 | 2026-05-07 | View source |
| LMArena | Coding | 54 | 1,486.2 | 1,644 | 2026-05-07 | View source |
| LMArena | Reasoning / hard prompts | 63 | 1,451.77 | 3,815 | 2026-05-07 | View source |
| LMArena | Text overall | 75 | 1,422.91 | 6,300 | 2026-05-07 | View source |
Public activity counters
Cumulative Hugging Face model-card downloads / likes / spaces
| Platform | Entry | Downloads | Likes | Spaces | Time | Source |
|---|---|---|---|---|---|---|
| huggingface | XiaomiMiMo/MiMo-V2.5 | 220,841 | 351 | 0 | 2026-07-15T20:11:41.551Z | View source |
Operations
Runtime / access signals
| Time | Signal | Original text | Source |
|---|---|---|---|
| 07/06/2026, 04:09 AM UTC | Need update parser in vllm | When i use tool calling, reasoning content was parsed uncorrectly, remain main content with <think token Plz Update like ke this PR (https://g... | Source |
| 06/19/2026, 05:12 PM UTC | Does the developers know this model has thought loop issues? | I'd really like to use this model as my primary model, but the looping issue prevents me from using it. I found a prompt which will make the m... | Source |
| 05/06/2026, 05:26 PM UTC | Roo Code: MiMo-V2.5 chat template crashes when system.content is an OpenAI content-part array | Hi Xiaomi MiMo team, I found a compatibility issue when serving XiaomiMiMo/MiMo-V2.5 with the recommended SGLang image and using OpenAI-compat... | Source |
Evidence gaps
Public display boundary
No linkable real reviewSpeed/latency not disclosed