xiaomi
Xiaomi: MiMo-V2.5-Pro
xiaomi/mimo-v2.5-pro
Our take
AI-assisted analysis based on the sources below. Suitability is an editorial assessment; test it on your own workload.
Xiaomi MiMo-V2.5-Pro is a text-only, open-source mixture-of-experts model with 1.02 trillion total and 42 billion active parameters, a hybrid attention stack, three multi-token-prediction layers, and up to a 1M-token context. Xiaomi positions it for agentic work, complex software engineering, and long-horizon tasks; vendor-reported GraphWalks testing indicates useful long-context behavior that still declines near the limit. OpenRouter lists tool calling, structured outputs, and optional reasoning controls. Evidence comes from a vendor card plus catalog metadata, so workload validation remains important.[1][2]
Strengths
- Open-source MoE design: 1.02T total parameters, 42B active, hybrid sliding-window/global attention, and three MTP layers.[2]
- Vendor-reported GraphWalks results indicate the model still performs both subtasks at very long inputs, though accuracy falls near 1M tokens.[2]
- OpenRouter catalog lists tools, structured outputs, response formatting, logprobs, sampling controls, and optional reasoning parameters.[1]
Limitations
- Only text input and output are documented; no multimodal capability is evidenced.[1][2]
- The pack contains a vendor card plus one aggregator catalog; independent results are limited to catalog-reported fields, so task-level verification is needed.[1][2]
- No knowledge cutoff or concrete hardware requirements are given; the card's serving examples are distributed deployments.[1][2]
Where to use it
- Consider for long-document analysis or repository-scale software tasks where a very large text context may help; test recall at required lengths.[2]
- Evaluate for tool-using agents or structured-output coding assistants, based on catalog-listed interfaces rather than proven outcomes.[1]
- Potentially suitable for English or Chinese text workflows, the only languages named in the card metadata.[2]
Sources
- [1] OpenRouter catalog · Xiaomi: MiMo-V2.5-ProSources checked: Nov 13, 2026
- [2] Model documentation · Xiaomi: MiMo-V2.5-ProSources checked: Nov 13, 2026
About this model
Description below is the original text supplied by OpenRouter, shown in the source language.
MiMo-V2.5-Pro is Xiaomi’s flagship model, delivering strong performance in general agentic capabilities, complex software engineering, and long-horizon tasks, with top rankings on benchmarks such as ClawEval, GDPVal, and SWE-bench Pro....
Specifications
- Model ID
- xiaomi/mimo-v2.5-pro
- Canonical ID
- xiaomi/mimo-v2.5-pro-20260422
- Creator
- xiaomi
- Added to OpenRouter
- 22.04.2026
- Context length
- 1,024,000
- Modality
- text->text
- Input modalities
- text
- Output modalities
- text
- Token pricing
- USD per 1M tokens: in $0.435 / out $0.870
- Knowledge cutoff
- —
- Expiration date
- —
Tasks and capabilities
A grade appears only where the source supplies a matching measurement for that task, the model is active and the cohort holds at least 10 comparable models. The capability column lists facts the source declares — supported formats, reasoning and tool parameters — which are not measured quality.
| Task | Grade | Source-declared capability |
|---|---|---|
| General quality | No measurement | reasoning: optional · parameters: include_reasoning, reasoning |
| Coding | BB — next 30% of the cohort | — |
| Agents | BB — next 30% of the cohort | parameters: structured_outputs, tool_choice, tools |
| Writing | No measurement | output: text |
| Reasoning | No measurement | reasoning: optional · parameters: include_reasoning, reasoning |
| Document analysis | No measurement | — |
| Vision | No measurement | — |
| Image generation | No measurement | — |
| Audio | No measurement | — |
| Video | No measurement | — |
| Embeddings | No measurement | — |
Evaluations
| Source | Benchmark | Value | Grade | Data retrieved |
|---|---|---|---|---|
| Artificial Analysis | Coding index | 60.2 | BB — next 30% of the cohort | 13.11.2026 |
| Artificial Analysis | Agentic index | 21.3 | BB — next 30% of the cohort | 13.11.2026 |
| Design Arena | models · 3dwin rate 54.1% · rank 38 | 1,235Elo | BB — next 30% of the cohort | 13.11.2026 |
| Design Arena | models · asciiartwin rate 46.6% · rank 42 | 1,151Elo | BB — next 30% of the cohort | 13.11.2026 |
| Design Arena | models · codecategorieswin rate 52.2% · rank 35 | 1,262Elo | AA — next 20% of the cohort | 13.11.2026 |
| Design Arena | models · datavizwin rate 51% · rank 32 | 1,258Elo | AA — next 20% of the cohort | 13.11.2026 |
| Design Arena | models · gamedevwin rate 54% · rank 28 | 1,265Elo | AA — next 20% of the cohort | 13.11.2026 |
| Design Arena | models · svgwin rate 48.8% · rank 37 | 1,184Elo | BB — next 30% of the cohort | 13.11.2026 |
| Design Arena | models · uicomponentwin rate 51% · rank 46 | 1,244Elo | BB — next 30% of the cohort | 13.11.2026 |
| Design Arena | models · websitewin rate 51.4% · rank 33 | 1,263Elo | AA — next 20% of the cohort | 13.11.2026 |
How grades work
Each letter grade compares a model only against other models scored by the same source and metric in the current snapshot. S means the top 10%, A the next 20%, B the next 30%, C the rest. Models without a published measurement are shown as not rated — a supported format alone never implies measured quality.
- S — S — top 10% of the cohort
- A — A — next 20% of the cohort
- B — B — next 30% of the cohort
- C — C — below the B band
Providers
Optional live provider statistics from OpenRouter. Missing values are unknown, not zero.
| Provider | Context length | Price in /1M | Price out /1M | Latency p50 (30m, s) | Throughput p50 (30m, tok/s) | Uptime (30m) |
|---|---|---|---|---|---|---|
| AtlasCloudfp8 | 1,024,000 | $0.435 | $0.870 | — | — | 100% |
| DigitalOcean | 262,144 | $0.480 | $1.8 | — | — | 100% |
Provider data retrieved: 13.11.2026
Related articles
No published articles identify this model yet.