MiMo-V2-Pro is Xiaomi's flagship foundation model released in March 2026, featuring over 1 trillion total parameters with 42 billion active, designed to serve as the brain of agent systems. It supports a 1M-token context window and is deeply optimized for agentic scenarios with strong tool-call generation, multi-step reasoning, and high schema adherence for structured JSON outputs. On PinchBench and ClawEval agent benchmarks, it achieves globally leading results, with coding ability surpassing Claude Sonnet 4.6 and general agent performance approaching Opus 4.6.
API|Reasoning|Proprietary Model
Knowledge Cutoff
Unknown
Input → Output Format
Context Memory
1.0MIN131KOUT
AI Performance Evaluation
Arena Overall Score
1449
±6As of 2026-05-01
Overall Rank
No.42
14,085 Votes
Arena by Ability
Hard Prompts
1477±7No.36
Expert Knowledge
1500±18No.15
Instruction Following
1450±9No.31
Conversation Memory
1467±12No.34
Creative
1422±13No.44
Coding
1506±10No.31
Math
1456±18No.31
Arena by Occupation
Creative Writing
1437±10No.35
Social Sciences
1482±12No.25
Media
1427±12No.37
Business
1460±12No.28
Healthcare
1477±19No.32
Legal
1455±19No.51
Software
1491±8No.35
Mathematics
1469±21No.24
Source:Arena Intelligence
Overall
AA Intelligence Index
49%↑10%
LiveBench
58%↓2%
Reasoning & Math
GPQA Diamond
87%↑5%
HLE
28%↑11%
LB Reasoning
70%↑1%
LB Math
77%↑3%
LB Data
49%↓4%
Coding
AA Coding Index
41%↑5%
LB Coding
69%↓4%
LB Agentic
30%↓15%
TAU2
95%↑15%
TerminalBench
41%↑7%
SciCode
43%↑1%
Language & Instructions
IFBench
69%↑6%
AA-LCR
61%↓1%
LB Language
69%↓3%
LB IF
43%↓8%
Output Speed
Standard Mode
63tok/s↓14
First Output 48.41s
