MiMo v2.5 Pro
MiMo v2.5 ProAlternatives?
CheaperGLM 5.3 FlashunstatedOpensame-56%$0.15 / $0.50Defaults to max · Self-hosted+4
BetterGPT-5.6 Solxhighwins 52%+1,371%$1.18OpenGLM 5.3maxwins 55%+295%$2.01Reasoning cannot be disabled+3CheaperMiMo v2.6 FlashunstatedOpensame-68%$0.14 / $0.28Pass back reasoning_content+3CheaperGPT-5.6 Lunaxhighsame-17%$0.085GLM 5.3 FlashunstatedOpensame-56%$0.15 / $0.50Defaults to max · Self-hosted+4MiMo v2.6 FlashunstatedOpensame-68%$0.14 / $0.28Pass back reasoning_content+3GPT-5.6 Lunaxhighsame-17%$0.085
Hy3unstatedOpensame-58%$0.13 / $0.53Context: 1M → 256K
Qwen3 235B A22B Instruct 2507unstatedOpenwins 46%-62%$0.09 / $0.55Context: 1M → 262KGPT-5.6 Solxhighwins 52%+1,371%$1.18
Gemini 3.8 Flashhighwins 56%+176%*$1.24Sampling parameters are ignored+1Gemini 3.6 Flashhighwins 53%+176%*$1.60
Muse Spark 1.3maxwins 59%+268%$1.60Only auto tool_choice is supported+5
Kimi K3maxOpenwins 54%+1,003%$2.00Memory (4-bit): 614 GB → 1,668 GB+4GLM 5.3maxOpenwins 55%+295%$2.01Reasoning cannot be disabled+3
NVIDIA Nemotron 3 Ultra 550B A55B NVFP4Opensame——Context: 1M → 262KGLM 5.2maxOpensame+295%$1.47GLM 5.1unstatedOpensame+295%$1.40 / $4.40Context: 1M → 203KGLM 5.3 FlashunstatedOpensame-56%$0.15 / $0.50Defaults to max · Self-hosted+4MiMo v2.6 FlashunstatedOpensame—$0.14 / $0.28Pass back reasoning_content+3MiMo v2.5unstatedOpenwins 45%—$0.14 / $0.28MiMo v2 Prounstatedwins 44%—$0.43 / $0.87MiMo v2 Omniunstatedwins 41%—$0.14 / $0.28Context: 1M → 262KMiMo v2.6 ProthinkingOpenn/a—$0.13Memory (4-bit): 614 GB → 615 GB+4MiMo v2 Omni 0327n/a——Context: 1M → 256KMiMo 7B BaseOpenn/a——Context: 1M → 33KMiMo 7B RLOpenn/a——Context: 1M → 33KAll Xiaomi models →Run it yourself
estimateHistory
3 versionsBenchmarks
in Chat & KnowledgeFAQ
How much does MiMo v2.5 Pro cost?›
$0.43 per 1M input tokens and $0.87 per 1M output tokens; cached input $0.004 per 1M. A typical benchmark task costs about $0.054 at default effort.
Is MiMo v2.5 Pro free to use?›
There is no free API; the cheapest paid price is $0.43 per 1M input tokens. The weights are open, so you can run it yourself.
How good is MiMo v2.5 Pro, and what is it best at?›
#59 of 334 in Chat & Knowledge, #75 of 132 in Coding, #71 of 133 in Agents & Tool Use, #42 of 195 in Writing & Language, #91 of 133 in Vision & Documents. Strongest at: Tool calling, Reasoning, Science, Long context, Knowledge, Instructions.
Is MiMo v2.5 Pro better than MiMo v2 Pro?›
On Arena Text · German, MiMo v2.5 Pro scores 1,458.9 vs 1,417.8 — higher.
How can I use MiMo v2.5 Pro?›
Through the API of AtlasCloud, DeepInfra, Novita, Xiaomi.
What is the context window of MiMo v2.5 Pro?›
1M tokens, with up to 131K output tokens.
When was MiMo v2.5 Pro released? Is there a newer version?›
Released 2026-04-22. It has been superseded by MiMo v2.6 Pro.
Is MiMo v2.5 Pro open source?›
Yes, the weights are open (MIT licence).
Can I run MiMo v2.5 Pro locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can MiMo v2.5 Pro read images?›
No, it accepts text only.
What is MiMo v2.5 Pro?›
MiMo v2.5 Pro is an open-weights language model from Xiaomi, released Apr 2026, the successor to MiMo v2 Pro.
What are the best alternatives to MiMo v2.5 Pro?›
GLM 5.3 Flash (cheaper), GPT-5.6 Sol (better), GLM 5.3 (open weights).