Model
Model explorer

MiMo-V2.5

OPEN
Xiaomi · MiMo family · released Apr 22, 2026

Xiaomi's native omnimodal flagship-tier model (text/image/video/audio), 310B total / 15B active MoE, 1M context, fully open-sourced under MIT. Consolidates the prior separate MiMo-V2-Pro (reasoning) and MiMo-V2-Omni (multimodal) lines into one architecture built on the MiMo-V2-Flash backbone.

ReasoningCodingVisionFunction callingTool useAgentic
2077.0
Elo · rank #55
Parameters
310B
Active params
15B (MoE)
Context
1M tokens
Architecture
Sparse MoE (256 routed experts, 8/token), hybrid Sliding-Window+Global attention 5:1, native 729M-param ViT + 261M-param audio encoder, 3x MTP modules
License
MIT
Languages
API price (in/out)
$0.14 / $0.28
Modalities
text · vision · audio · video
Benchmark results
Bar shows position within the tracked field; marker = field best
Agents' Last ExamAgents23.6%#16
best: GPT-5.6 · 52.7%
Arena EloHuman preference1434#12
best: Claude Fable 5 · 1505
CharXivVision81.0%#8
best: Muse Spark 1.1 (xhigh) · 88.4%
GDPval-AAAgents1145#29
best: Claude Fable 5 · 1932
GPQA DiamondReasoning84.9%#56
best: GPT-5.6 · 94.6%
Humanity's Last ExamReasoning25.2%#55
best: Claude Sonnet 5 · 57.4%
IFBenchReasoning67.1%#27
best: MiniMax M3 · 83.0%
LiveCodeBenchCoding81.5%#41
best: DeepSeek-V4-Pro (Think Max) · 93.5%
MMLU-ProKnowledge82.9%#40
best: Claude Fable 5 · 91.5%
MMMU-ProVision77.9%#17
best: Claude Opus 4.7 · 85.5%
PinchBenchAgents89.7%#3
best: Trinity-Large-Thinking · 91.9%
SWE-bench ProCoding56.1%#22
best: Claude Fable 5 · 80.0%
SWE-bench VerifiedCoding71.0%#57
best: Claude Fable 5 · 95.0%
τ²-Bench TelecomAgents90.6%#19
best: Claude Opus 4.6 · 99.3%
τ³-BankingAgents6.6%#4
best: Grok 4.5 · 33.0%
Terminal-Bench 2.0Coding65.8%#24
best: GPT-5.6 · 88.8%
Video-MMEVision87.7%#4
best: Seed 2.1 Pro · 89.2%
Run it locally
VRAM @ Q4
VRAM @ FP16
Fits on (Q4)
Multi-node cluster required
Throughput data unavailable.
Quantizations
MiMo family
Elo progression across releases
API price $0.14/$0.28 · each benchmark row carries its own source badge (see methodology)