Gemini 4 Argon
Claude Opus 5.5
Muse Wonder
Qwen3.8-Max
Kimi K3
GLM-5.3
GPT-5.5
ERNIE 5.1
DeepSeek-V4.1-Flash-20260910
Grok 4.20 Beta
amazon-nova-experimental-chat
Dola Seed 2.0 Pro
NVIDIA-Nemotron-3-Ultra-550B-A55B
Gemma 4 31B
Step 5 Preview
MiMo-V2.5
MiMo-V2.5 is an open-weight model from Xiaomi, released Apr 22, 2026. It ranks #100 of 351 on the LMArena text leaderboard with a rating of 1428. It costs $0.12 per million input tokens and $0.24 per million output tokens, with a 1.1M-token context window.
- Developer
- Xiaomi
- Released
- Apr 22, 2026
- Context window
- 1.1M
- Price in / out, per 1M tokens
- $0.12 / $0.24
- Weights
- Open · MIT
- Inputs
- Text, Audio, Images, Video
- Reasoning
- Yes
- Tool use
- Yes
- Hugging Face downloads, 30 days
- 143.8K
- Hugging Face likes
- 431
The rating LMArena published at each update since Apr 28, 2026. Ratings move as votes come in and new models join.
Benchmarks
| Benchmark | Score bar | Accuracy | Rank | Cost / test |
|---|---|---|---|---|
| SWE-benchCoding | 71.00% | 57 / 85 | $0.01 | |
| Vibe Code BenchCoding | 42.17% | 67 / 115 | $0.07 | |
| LiveCodeBenchCoding | 81.51% | 61 / 143 | $0.14 / $0.28 | |
| Tax Agent BenchFinance | 7.77% | 65 / 76 | $0.02 | |
| Legal Research BenchLaw | 9.13% | 77 / 85 | $0.03 | |
| MedCodeMedicine | 31.89% | 92 / 107 | $0.14 / $0.28 | |
| GPQA DiamondReasoning | 81.57% | 65 / 138 | $0.14 / $0.28 | |
| MMLU-ProReasoning | 82.93% | 77 / 138 | $0.14 / $0.28 | |
| MMMUReasoning | 80.00% | 45 / 93 | $0.14 / $0.28 |
Vals AI, as of Oct 24, 2026.
LMArena, by category
| Leaderboard | Rating | Rank | Votes |
|---|---|---|---|
| TextLMArena | 1428 | 100 / 351 | 47.4K |
| CodingLMArena | 1469 | 71 / 346 | 13.7K |
| MathLMArena | 1437 | 89 / 339 | 2.5K |
| Creative writingLMArena | 1393 | 112 / 349 | 7.9K |
| Instruction followingLMArena | 1435 | 68 / 351 | 16.1K |
| Hard promptsLMArena | 1450 | 74 / 351 | 31.9K |
| Multi-turnLMArena | 1445 | 77 / 349 | 8.4K |
| Long queriesLMArena | 1445 | 70 / 330 | 21.3K |
| ExpertLMArena | 1459 | 74 / 305 | 4.8K |
| Legal and governmentIndustries | 1425 | 122 / 325 | 3.8K |
| Medicine and healthcareIndustries | 1432 | 110 / 325 | 3.4K |
| Business and financeIndustries | 1432 | 75 / 344 | 9.7K |
| ChineseLanguages | 1469 | 88 / 328 | 2.6K |
| SpanishLanguages | 1414 | 126 / 249 | 1.4K |
| GermanLanguages | 1422 | 95 / 261 | 797 |
| FrenchLanguages | 1448 | 93 / 252 | 1.8K |
| JapaneseLanguages | 1306 | 158 / 236 | 540 |
| WebDevLMArena | 1438 | 82 / 124 | 14.9K |
| VisionLMArena | 1247 | 66 / 146 | 29.4K |
LMArena, CC BY 4.0, as of Oct 22, 2026. Where a model is rated at several effort levels, its best is shown.
Compare MiMo-V2.5
Xiaomi
Xiaomi’s MiMo models are open reasoning models built for devices and cloud. Founded 2010, based in Beijing. xiaomimimo.github.io
Questions
Who makes MiMo-V2.5?
MiMo-V2.5 is made by Xiaomi, founded 2010 and based in Beijing.
How much does MiMo-V2.5 cost?
$0.12 per million input tokens and $0.24 per million output tokens, at its provider’s list price.
What is MiMo-V2.5’s context window?
1.1M tokens (1,050,000).
How does MiMo-V2.5 rank?
#100 of 351 on LMArena’s text leaderboard, rated 1428 (95% interval 1424–1432) from 47,364 votes, as of Oct 22, 2026.
Is MiMo-V2.5 open source?
MiMo-V2.5 has open weights under the MIT license, downloadable from Hugging Face (XiaomiMiMo/MiMo-V2.5).