XiaomiMiMo-V2.5
MiMo-V2.5 is a Xiaomi open-weights reasoning model, released 22 Apr 2026. Its best configuration ranks #255 of 785 on the BenchLeader Index at 54.3 ±7.0, in the upper half. The ± is the point: 371 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in instruction following (66) and lowest in reasoning (47). At $0.165 per million tokens blended it is among the cheapest fifth of ranked models. Output speed of 76 tokens per second puts it slower than most, with a first answer in 27.4 s. Last measured 22 Apr 2026.stale: no new result in six months
- Blended price
- $0.165/M
- $0.130 in · $0.270 out
- Output speed
- 76 tok/s
- measured by Artificial Analysis
- First answer
- 27 s
- Context
- 1M
- Full answer
- 34 s
- median, reasoning included
- Cost per run
- $0.038
- one full Intelligence Index run
- Released
- 22 Apr 2026
- Overall index54
- Reasoning47
- Agents & tools48
- Knowledge57
- Instruction following66
- Multimodal58
- Long context61
- Composite54
How its index has moved
8 Sept 2026 to 21 Oct 2026The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.
Benchmark results
See all LLM benchmarks. One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Reasoning
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| GPQA Diamond (AA)not in index | 85.0%#130 | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 27.2%#176 | Artificial Analysis | |
| CritPt | 3.7%#171 | Artificial Analysis |
Coding
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| SciCode (AA)not in index | 43.9%#152 | Artificial Analysis | |
| Terminal-Bench 4.0 (AA)not in index | 0.0%#152 | Artificial Analysis | |
| Terminal-Bench 2.1 (AA)not in index | 63.7%#88 | Artificial Analysis |
Agents & tools
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Terminal-Bench Hard | 41.7%#50 | Artificial Analysis | |
| τ²-Bench Telecom (AA)not in index | 90.6%#61 | Artificial Analysis | |
| τ³-Banking (AA) | 8.7%#155 | Artificial Analysis | |
| AutomationBenchnot in index | 18.4%#139 | Artificial Analysis | |
| GDP.pdf (AA)not in index | 4.0%#170 | Artificial Analysis |
Knowledge
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| AA-Omniscience | -9.8#190 | Artificial Analysis | |
| AA-Omniscience: accuracynot in index | 16.8%#381 | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | 68.1%#51 | Artificial Analysis |
Instruction following
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| IFBench | 67.1%#89 | Artificial Analysis |
Multimodal
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| MMMU-Pro | 75.4%#99 | Artificial Analysis |
Long context
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| AA-LCR | 73.0%#182 | Artificial Analysis |
Composite
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| AA Intelligence Index v4.3.2 | 21.8#208 | Artificial Analysis |
Price history
The price above, per 1M tokens, every day it changed.
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0001 | – | 31.3 s |
| Summarize a 30-page report | 12,000 / 600 | $0.0017 | – | 35.2 s |
| Code edit | 6,000 / 1,500 | $0.0012 | – | 47.0 s |
| Agentic coding session | 60,000 / 4,000 | $0.0089 | – | 1.3 min |
| Structured extraction | 2,000 / 200 | $0.0003 | – | 30.0 s |
See also
Questions people ask
- What are MiMo-V2.5's benchmark scores?
- As of 24 Oct 2026, MiMo-V2.5 scores 54.3 ±7.0 on the BenchLeader Index, strongest in instruction following (66) and weakest in reasoning (47). That rests on 8 independent benchmark results, and every score on this page links to its publisher.
- How does MiMo-V2.5 rank?
- MiMo-V2.5 is #255 of 785 ranked configurations on 24 Oct 2026. 371 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Opus 5.5 (max effort) at 72.6.
- How much does MiMo-V2.5 cost per 1M tokens?
- As of 24 Oct 2026, MiMo-V2.5 costs $0.130 per 1M input tokens and $0.270 per 1M output tokens, $0.165 blended (three input tokens per output token). The list price last changed on 23 Oct 2026.
- When was MiMo-V2.5 released?
- Xiaomi released MiMo-V2.5 on 22 Apr 2026.
Data as of 24 Oct 2026. Compare with another model.
Cite as: BenchLeader, “MiMo-V2.5: benchmarks, pricing, speed and rank”, data as of 24 Oct 2026, https://www.benchleader.com/models/mimo-v2-5.