benchgap
Xiaomi · model

MiMo-V2.6-Flash benchmark scores

As of 2026-11-21, MiMo-V2.6-Flash (Xiaomi) has measured scores on 24 benchmarks and estimated scores on 131 more.

BenchmarkScoreSource
AA-GPQA Diamond89.7%estimated ± 9.2 pp, medium confidence
AA-HLE35.1%measured
AA-SciCode51.3%measured
Artificial Analysis Intelligence Index26.7%measured
CritPt12.0%measured
GDPval-AA55.4%measured
GPQA Diamond89.9%estimated ± 2.4 pp, high confidence
HLE38.2%estimated ± 8.5 pp, medium confidence
AA-LCR74.3%measured
LiveCodeBench v691.8%estimated ± 4.9 pp, high confidence
AA-Omniscience Accuracy27.2%measured
SciCode51.1%estimated ± 3.1 pp, high confidence
Terminal-Bench 2.187.6%measured
BrowseComp83.3%estimated ± 3.0 pp, medium confidence
HLE w/ tools63.8%estimated ± 4.9 pp, low confidence
LongBench v261.2%estimated ± 5.5 pp, medium confidence
AA Coding Index75.5%estimated ± 2.5 pp, high confidence
AA-MMMU-Pro73.1%measured
APEX-Agents64.1%estimated ± 7.4 pp, low confidence
APEX-Agents-AA41.5%estimated ± 3.6 pp, low confidence
BioMysteryBench (human-difficult)45.1%estimated ± 6.8 pp, low confidence
DeepSearchQA86.7%estimated ± 7.9 pp, low confidence
SWE-bench Verified87.1%estimated ± 5.1 pp, medium confidence
AutomationBench52.3%measured
BFCL v477.0%estimated ± 2.7 pp, medium confidence
CyberGym95.1%measured
JobBench61.2%measured
SWE-bench Pro60.8%estimated ± 6.3 pp, medium confidence
WideResearch83.7%estimated ± 1.2 pp, medium confidence
AA Agentic Index51.4%estimated ± 3.1 pp, high confidence
MMLU-Pro85.2%estimated ± 5.0 pp, medium confidence
GPQA89.8%estimated ± 7.8 pp, medium confidence
MMLU90.5%estimated ± 0.9 pp, high confidence
Vals GPQA Diamond88.0%estimated ± 5.7 pp, medium confidence
Vals LiveCodeBench84.2%estimated ± 5.7 pp, low confidence
Vals MMLU-Pro83.8%estimated ± 2.9 pp, high confidence
MMMLU89.8%estimated ± 0.7 pp, medium confidence
MMMU82.2%estimated ± 2.7 pp, high confidence
Chess Puzzles39.3%estimated ± 8.0 pp, low confidence
AA MMLU-Pro85.2%estimated ± 0.8 pp, high confidence
Gert Labs73.5%estimated ± 5.5 pp, low confidence
AA-AnalystAgent41.7%estimated ± 5.5 pp, low confidence
AA EnterpriseOps-Gym46.7%estimated ± 1.2 pp, medium confidence
AA Tau3 Banking50.2%estimated ± 3.3 pp, low confidence
AA Terminal-Bench 2.184.8%estimated ± 2.5 pp, high confidence
ApprenticeBench35.5%estimated ± 4.0 pp, medium confidence
ARC-AGI-183.7%estimated ± 7.8 pp, low confidence
ARC-AGI-253.6%estimated ± 14.4 pp, low confidence
CursorBench 3.147.2%estimated ± 3.4 pp, low confidence
CursorBench 3.227.1%estimated ± 3.6 pp, low confidence
EBR-bench32.3%estimated ± 9.9 pp, medium confidence
Terminal-Bench 3.017.7%estimated ± 3.2 pp, high confidence
FrontierCode 1.1 Main48.7%estimated ± 3.5 pp, low confidence
FrontierSWE v221.0%estimated ± 14.2 pp, low confidence
Furniture Assembly32.1%estimated ± 6.5 pp, low confidence
MaxxBench56.2%estimated ± 5.7 pp, medium confidence
MazeBench0.2%estimated ± 3.7 pp, high confidence
MirrorCode34.5%estimated ± 14.9 pp, low confidence
Mystery Game Puzzles12.4%estimated ± 7.5 pp, low confidence
NYT Connections100.0%estimated ± 8.9 pp, low confidence
OfficeQA Pro50.3%estimated ± 6.8 pp, low confidence
OSWorld-Verified80.8%measured
ParseBench-157.0%estimated ± 6.2 pp, medium confidence
Terminal-Bench 4.116.1%measured
Terminal-Bench Hard56.9%estimated ± 2.8 pp, medium confidence
Vals SWE-bench80.2%estimated ± 8.2 pp, medium confidence
Terminal-Bench 2.1 (Vals AI)76.4%measured
CWE-Bench-AA61.0%estimated ± 6.0 pp, low confidence
AA Cyber Index46.3%estimated ± 9.4 pp, medium confidence
DeepsecBench-AA18.7%estimated ± 5.2 pp, low confidence
HealthBench Professional (length-adjusted, Artificial Analysis)54.0%estimated ± 4.3 pp, low confidence
HealthBench Professional (raw, Artificial Analysis)58.2%estimated ± 0.8 pp, low confidence
AA ITBench43.9%estimated ± 4.7 pp, low confidence
Bug Hunt Bench29.7%estimated ± 10.4 pp, low confidence
Chartography (tools)79.0%estimated ± 4.1 pp, low confidence
CursorBench 4.046.2%estimated ± 4.7 pp, high confidence
CWE-bench v148.9%estimated ± 4.5 pp, low confidence
DeepSWE67.9%measured
DRACO85.4%estimated ± 2.0 pp, medium confidence
FrontierCode 2.026.1%estimated ± 9.3 pp, low confidence
FrontierPhysics0.9%estimated ± 2.7 pp, medium confidence
HLE w/o tools37.3%estimated ± 4.9 pp, high confidence
OSWorld 2.015.5%estimated ± 3.6 pp, high confidence
OSWorld 2.170.4%estimated ± 6.0 pp, medium confidence
OSWorld 2.1 Offline71.7%estimated ± 1.9 pp, medium confidence
PostTrainBench v1.135.8%estimated ± 4.5 pp, high confidence
ProgramBench26.0%measured
SpireClimb1.4%estimated ± 5.4 pp, low confidence
StealthMark49.3%measured
Terminal-Bench 4.028.8%measured
Terminal-Bench-Science 0.10.0%estimated ± 10.7 pp, low confidence
Toolathlon-Verified73.6%measured
VulcanBench Frontier v488.5%estimated ± 1.2 pp, low confidence
AA AutomationBench40.6%estimated ± 4.8 pp, low confidence
GDP.pdf18.7%estimated ± 4.1 pp, medium confidence
AA Terminal-Bench 4.015.7%estimated ± 6.2 pp, medium confidence
ARC-AGI-31.3%estimated ± 9.8 pp, medium confidence
ARC-AGI-3 Provider Adapter2.7%estimated ± 7.3 pp, low confidence
cursorBench4144.8%estimated ± 3.7 pp, medium confidence
CWE-bench45.0%estimated ± 4.6 pp, medium confidence
MKA24.4%estimated ± 3.1 pp, medium confidence
FACTS Multimodal34.6%estimated ± 2.7 pp, low confidence
Hermes Bench65.7%estimated ± 1.5 pp, low confidence
Hermes Index21.5%estimated ± 5.5 pp, low confidence
ParseBench-262.0%estimated ± 6.2 pp, medium confidence
CharXiv79.7%estimated ± 6.3 pp, low confidence
CharXiv w/o tools79.0%estimated ± 2.5 pp, medium confidence
ExploitBench25.3%measured
ExploitGym6.0%measured
AI2D_TEST89.2%estimated ± 2.0 pp, medium confidence
C-Eval93.6%estimated ± 1.0 pp, medium confidence
Claw-Eval76.7%estimated ± 4.5 pp, medium confidence
CountBench93.7%estimated ± 2.9 pp, medium confidence
DynaMath84.4%estimated ± 3.0 pp, medium confidence
ERQA54.9%estimated ± 3.1 pp, high confidence
MathVision82.7%estimated ± 3.1 pp, high confidence
MCP Atlas82.2%estimated ± 7.5 pp, medium confidence
MMLU-Redux94.4%estimated ± 0.8 pp, medium confidence
MMMU-Pro74.4%estimated ± 1.5 pp, high confidence
NL2Repo59.9%estimated ± 6.0 pp, medium confidence
RealWorldQA81.8%estimated ± 2.9 pp, medium confidence
ScreenSpot Pro73.3%estimated ± 10.8 pp, low confidence
SuperGPQA66.6%estimated ± 9.3 pp, medium confidence
Terminal-Bench 2.077.0%estimated ± 5.8 pp, medium confidence
Toolathlon55.9%estimated ± 6.3 pp, medium confidence
V*100.0%estimated ± 7.5 pp, medium confidence
Video-MME (w/o subtitle)81.7%estimated ± 1.7 pp, medium confidence
Video-MME (with subtitle)83.8%estimated ± 3.7 pp, medium confidence
HealthBench Hard30.0%estimated ± 8.6 pp, low confidence
MedXpertQA (MM)66.8%estimated ± 6.5 pp, low confidence
React Native Evals77.2%estimated ± 4.7 pp, medium confidence
ResearchClawBench19.4%estimated ± 2.1 pp, medium confidence
DAYJOB0.7%estimated ± 1.1 pp, low confidence
DAYJOB Healthcare1.5%estimated ± 1.5 pp, low confidence
BioMysteryBench (human-solvable)85.1%estimated ± 0.6 pp, low confidence
HealthBench (raw)44.1%estimated ± 6.0 pp, low confidence
HealthBench (length-adjusted)51.7%estimated ± 3.4 pp, low confidence
HealthBench Professional56.8%estimated ± 4.2 pp, low confidence
HealthBench Professional (raw)54.9%estimated ± 4.5 pp, medium confidence
HLE-Verified0.0%estimated ± 6.8 pp, low confidence
LABBench277.7%estimated ± 2.4 pp, low confidence
ScarfBench65.7%estimated ± 3.8 pp, medium confidence
STEMBench2.4%estimated ± 0.2 pp, medium confidence
SimpleQA28.0%estimated ± 4.3 pp, medium confidence
Agents' Last Exam27.6%measured
BabyVision w/ Python86.4%estimated ± 4.0 pp, low confidence
OpenHarmony Bench v1.058.1%estimated ± 2.4 pp, high confidence
SEC-Bench Pro47.5%measured
MRCRv283.6%estimated ± 11.3 pp, low confidence
MMLongBench-Doc61.2%estimated ± 2.6 pp, low confidence
SWE-Marathon42.9%estimated ± 12.9 pp, low confidence
BioBasicsBench33.9%estimated ± 2.8 pp, low confidence
MMMU-Pro w/ Python76.5%estimated ± 2.0 pp, high confidence
ExploitBench Internal Port0.0%estimated ± 6.5 pp, medium confidence
SWE-Atlas Codebase QnA60.0%estimated ± 4.8 pp, high confidence