benchgap
Xiaomi · model

MiMo-V2.5 benchmark scores

As of 2026-11-22, MiMo-V2.5 (Xiaomi) has measured scores on 13 benchmarks and estimated scores on 105 more.

BenchmarkScoreSource
AA-GPQA Diamond82.5%estimated ± 3.6 pp, high confidence
AA-HLE20.2%estimated ± 6.2 pp, medium confidence
AA-SciCode48.3%estimated ± 3.6 pp, high confidence
Artificial Analysis Intelligence Index19.6%estimated ± 5.8 pp, medium confidence
CritPt2.0%estimated ± 5.3 pp, medium confidence
GDPval-AA32.8%estimated ± 8.0 pp, low confidence
GPQA Diamond86.0%estimated ± 1.8 pp, medium confidence
HLE32.8%estimated ± 9.2 pp, low confidence
LiveCodeBench v687.1%estimated ± 2.2 pp, medium confidence
AA-Omniscience Accuracy28.9%estimated ± 6.5 pp, medium confidence
SciCode39.7%estimated ± 2.9 pp, low confidence
Terminal-Bench 2.178.5%estimated ± 6.7 pp, medium confidence
BrowseComp81.3%estimated ± 5.6 pp, medium confidence
HLE w/ tools48.5%estimated ± 5.7 pp, medium confidence
VITA-Bench39.6%estimated ± 10.7 pp, low confidence
AA Coding Index53.9%estimated ± 5.6 pp, medium confidence
AA-MMMU-Pro77.0%estimated ± 1.5 pp, high confidence
APEX-Agents-AA14.7%estimated ± 9.3 pp, medium confidence
DeepSearchQA80.5%estimated ± 10.2 pp, low confidence
SWE-bench Verified77.6%estimated ± 3.6 pp, high confidence
CyberGym54.2%estimated ± 11.0 pp, low confidence
JobBench32.5%estimated ± 9.3 pp, low confidence
SWE-bench Pro56.1%measured
WideResearch70.9%estimated ± 3.3 pp, high confidence
AA Agentic Index23.6%estimated ± 6.4 pp, medium confidence
MMLU-Pro83.7%estimated ± 4.8 pp, high confidence
GPQA83.7%estimated ± 4.5 pp, high confidence
MMLU91.2%estimated ± 1.2 pp, high confidence
Vals GPQA Diamond81.6%measured
Vals LiveCodeBench81.5%measured
Vals MMLU-Pro82.9%measured
MMMU84.9%estimated ± 1.7 pp, high confidence
AA MMLU-Pro85.6%estimated ± 1.1 pp, high confidence
Gert Labs29.9%measured
AA Terminal-Bench 2.180.8%estimated ± 2.5 pp, high confidence
ApprenticeBench3.3%estimated ± 10.4 pp, low confidence
ARC-AGI-171.1%estimated ± 6.5 pp, medium confidence
ARC-AGI-222.4%estimated ± 13.8 pp, low confidence
CursorBench 3.152.0%estimated ± 5.6 pp, low confidence
CursorBench 3.252.3%estimated ± 3.6 pp, high confidence
Terminal-Bench 3.06.1%estimated ± 6.9 pp, low confidence
FrontierCode 1.1 Main37.6%estimated ± 5.0 pp, high confidence
FrontierSWE v28.8%estimated ± 8.3 pp, medium confidence
MazeBench0.0%estimated ± 2.7 pp, high confidence
OSWorld-Verified70.4%estimated ± 3.8 pp, medium confidence
ParseBench-161.2%estimated ± 8.4 pp, medium confidence
Terminal-Bench 4.112.2%estimated ± 14.9 pp, low confidence
Terminal-Bench Hard48.1%estimated ± 3.9 pp, high confidence
Vals SWE-bench71.0%measured
Terminal-Bench 2.1 (Vals AI)60.7%measured
VulcanBench v381.6%estimated ± 5.6 pp, low confidence
CWE-Bench-AA40.0%estimated ± 8.1 pp, medium confidence
AA Cyber Index20.3%estimated ± 12.4 pp, low confidence
DeepsecBench-AA9.2%estimated ± 8.2 pp, low confidence
AA ITBench36.8%estimated ± 4.6 pp, medium confidence
Bug Hunt Bench29.5%estimated ± 12.1 pp, low confidence
CursorBench 4.032.6%estimated ± 3.3 pp, medium confidence
CWE-bench v139.3%estimated ± 5.7 pp, medium confidence
DeepSWE56.3%estimated ± 7.2 pp, low confidence
FrontierCode 2.020.1%estimated ± 6.7 pp, low confidence
HLE w/o tools26.5%estimated ± 4.9 pp, high confidence
OSWorld 2.04.6%estimated ± 6.9 pp, low confidence
OSWorld 2.111.5%estimated ± 3.6 pp, medium confidence
OSWorld 2.1 Offline39.9%estimated ± 10.1 pp, low confidence
PostTrainBench v1.116.2%estimated ± 3.7 pp, medium confidence
Toolathlon-Verified63.3%estimated ± 3.0 pp, high confidence
VulcanBench Frontier v477.4%estimated ± 8.3 pp, low confidence
AA AutomationBench43.2%estimated ± 9.9 pp, low confidence
GDP.pdf13.4%estimated ± 4.2 pp, medium confidence
AA Terminal-Bench 4.010.2%estimated ± 13.7 pp, low confidence
ARC-AGI-30.0%estimated ± 8.6 pp, low confidence
CWE-bench33.3%estimated ± 6.2 pp, low confidence
MKA32.9%estimated ± 4.6 pp, medium confidence
Hermes Index20.4%estimated ± 5.7 pp, low confidence
ParseBench-269.0%estimated ± 5.0 pp, medium confidence
CharXiv81.0%measured
CharXiv w/o tools82.0%estimated ± 1.2 pp, high confidence
ExploitGym0.0%estimated ± 8.3 pp, low confidence
AI2D_TEST93.4%estimated ± 0.9 pp, medium confidence
Claw-Eval62.3%measured
CountBench96.7%estimated ± 2.0 pp, medium confidence
DynaMath86.0%estimated ± 0.8 pp, medium confidence
ERQA66.7%estimated ± 2.9 pp, high confidence
MathVision87.2%estimated ± 3.0 pp, high confidence
MCP Atlas64.6%estimated ± 12.9 pp, low confidence
MCP-Tasks77.5%estimated ± 6.4 pp, low confidence
MLVU (M-Avg)86.0%estimated ± 1.5 pp, high confidence
MMMU-Pro77.9%measured
NL2Repo40.5%estimated ± 4.7 pp, high confidence
PostTrainBench v113.2%estimated ± 3.6 pp, medium confidence
RealWorldQA83.8%estimated ± 1.5 pp, high confidence
ScreenSpot Pro73.2%estimated ± 6.8 pp, medium confidence
SimpleVQA66.2%estimated ± 7.5 pp, medium confidence
Terminal-Bench 2.065.8%measured
Toolathlon52.3%estimated ± 6.9 pp, medium confidence
Video-MME (w/o subtitle)84.4%estimated ± 1.3 pp, high confidence
Video-MME (with subtitle)87.7%measured
VideoMMMU85.4%estimated ± 1.5 pp, medium confidence
We-Math86.6%estimated ± 3.5 pp, medium confidence
AA LiveCodeBench87.1%estimated ± 2.6 pp, high confidence
LiveCodeBench Pro80.6%estimated ± 6.0 pp, low confidence
MedXpertQA (MM)70.0%estimated ± 2.4 pp, high confidence
React Native Evals77.3%estimated ± 4.7 pp, medium confidence
ResearchClawBench16.9%measured
PostTrainBench13.7%estimated ± 3.7 pp, medium confidence
Finance Agent v255.1%estimated ± 2.4 pp, low confidence
FrontierCode 1.1 Extended54.7%estimated ± 3.0 pp, low confidence
LABBench279.1%estimated ± 1.8 pp, low confidence
ScarfBench42.1%estimated ± 12.3 pp, low confidence
Agents' Last Exam29.7%estimated ± 9.7 pp, low confidence
BabyVision w/ Python88.8%estimated ± 3.8 pp, low confidence
OpenHarmony Bench v1.051.5%estimated ± 3.2 pp, high confidence
BabyVision60.5%estimated ± 9.1 pp, low confidence
MMVU77.1%estimated ± 2.7 pp, medium confidence
PerceptionBench59.6%estimated ± 4.9 pp, low confidence
OCRBench V271.1%estimated ± 3.3 pp, high confidence
LifeVal6.9%estimated ± 5.0 pp, low confidence
MMMU-Pro w/ Python80.5%estimated ± 2.6 pp, high confidence