MiMo-V2.5-Pro-FP4-DFlash
Model overview
MiMo-V2.5-Pro-FP4-DFlash
Xiaomi's official MiMo-V2.5-Pro FP4 and DFlash checkpoint for lower-latency local inference.
general LLMreasoningcodinglong contextchatcode generationsoftware engineering agentlocal inferenceevaluation
Representative version
MiMo-V2.5-Pro-FP4-DFlash
| Company | Xiaomi |
|---|---|
| Release | 2026-06-08 |
| Release type | Model checkpoint release |
| Parameters | 1,020B / 42B active |
| Architecture | Mixture-of-experts transformer with FP4 expert quantization and DFlash speculative decoding |
| Context | 1,000,000 |
| Input modalities | text / code |
| Output modalities | text / code |
| Tool use / function calling | Tool use: Not disclosed; Function calling: Not disclosed |
| Structured output / reasoning | Confirmed: Reasoning |
| Vision / audio / video | Vision: Not disclosed; Audio: Not disclosed; Video: Not disclosed |
| Open weights / License | Weights available / MIT License sourcexiaomi-mimo-v2-5-pro-fp4-dflash-current · 2026-08-04 |
| Vendor API | Not disclosed API source pending verification |
| OpenRouter | Not separately listed on OpenRouter |
| Pricing | No reliable public source |
| Benchmark | No reliable public source |
| Public signals | 1 community discussion / 1 activity snapshot |
| Use cases | Low-latency local inference / Coding and reasoning research / Speculative-decoding deployment / assistant chat |
| Claim evidence | Release event / Model specifications / Weight checkpointMiMo-V2.5-Pro-FP4-DFlash model cardMiMo-V2.5-Pro-FP4-DFlash · 2026-06-08 · MiMo-V2.5-Pro-FP4-DFlash · MiMo-V2.5-Pro-FP4-DFlash · MiMo-V2.5-Pro-FP4-DFlash |
OpenRouter
Pricing / context / availability
| OpenRouter | Not separately listed on OpenRouter |
|---|---|
| OpenRouter listing | Not separately listed in the current OpenRouter public list |
| Mapped version | MiMo-V2.5-Pro-FP4-DFlash xiaomi-mimo-v2-5-pro-fp4-dflash-current |
| Pricing | No reliable public source |
| Context | 1000000 |
| Speed/Latency | No reliable public source |
| Snapshot time | 2026-08-11T06:28:09.895Z |
| Source | View source |
Verified limitations
Usage boundaries
- Running a 1.02T-parameter MoE checkpoint still requires substantial accelerator memory and deployment expertise.
- No independently verified production latency or cost benchmark is attached.
Public activity counters
Hugging Face model-card activity snapshot
| Platform | Entry | Last-month downloads | Current likes | Current Spaces | Time | Source |
|---|---|---|---|---|---|---|
| huggingface Current HF activity snapshot | XiaomiMiMo/MiMo-V2.5-Pro-FP4-DFlash | 1,115 | 145 | 0 | View source |
Downloads are Hugging Face's trailing-month counter; likes and Spaces are snapshots at collection time.
Missing public information
Public information not yet available
Not listed on OpenRouterNo reliable benchmark sourceNo linkable model-specific reviewSpeed/latency not disclosed