benchgap
Alibaba (Qwen) · model

Qwen3.8-Omni-Flash benchmark scores

As of 2026-11-22, Qwen3.8-Omni-Flash (Alibaba (Qwen)) has measured scores on 16 benchmarks and estimated scores on 75 more.

BenchmarkScoreSource
AA-GPQA Diamond90.5%estimated ± 4.4 pp, high confidence
AA-HLE37.6%estimated ± 3.9 pp, high confidence
AA-SciCode49.7%estimated ± 3.1 pp, high confidence
Artificial Analysis Intelligence Index33.6%estimated ± 7.4 pp, medium confidence
CritPt11.7%estimated ± 4.0 pp, high confidence
GPQA Diamond91.0%measured
HLE36.5%measured
IFBench81.5%measured
LiveCodeBench v692.6%measured
AA-Omniscience Accuracy36.4%estimated ± 8.4 pp, medium confidence
SciCode56.3%estimated ± 4.9 pp, high confidence
IFEval94.9%estimated ± 3.6 pp, high confidence
AA Coding Index72.4%estimated ± 2.5 pp, high confidence
AA-MMMU-Pro80.1%estimated ± 1.9 pp, high confidence
SWE-bench Verified80.7%estimated ± 2.3 pp, high confidence
SWE-bench Pro63.3%measured
MMLU-Pro85.0%estimated ± 4.4 pp, high confidence
AA-IFBench80.4%estimated ± 7.4 pp, medium confidence
GPQA91.0%measured
MMLU88.1%estimated ± 4.2 pp, medium confidence
Vals GPQA Diamond90.1%estimated ± 2.1 pp, high confidence
Vals LiveCodeBench86.1%estimated ± 3.4 pp, low confidence
Vals MMLU-Pro87.2%estimated ± 1.6 pp, high confidence
MMMLU90.2%estimated ± 1.1 pp, medium confidence
MMMU85.8%estimated ± 3.7 pp, medium confidence
AA MMLU-Pro87.0%estimated ± 1.2 pp, low confidence
Gert Labs100.0%estimated ± 12.1 pp, low confidence
ARC-AGI-190.8%estimated ± 2.5 pp, high confidence
ARC-AGI-262.4%estimated ± 9.7 pp, medium confidence
CursorBench 3.156.8%estimated ± 5.6 pp, low confidence
CursorBench 3.262.0%estimated ± 3.6 pp, high confidence
FrontierCode 1.1 Main43.0%estimated ± 3.5 pp, low confidence
FrontierSWE v221.3%estimated ± 8.3 pp, medium confidence
ParseBench-168.3%estimated ± 1.0 pp, medium confidence
Vals SWE-bench82.6%estimated ± 5.2 pp, medium confidence
Bug Hunt Bench15.5%estimated ± 7.2 pp, low confidence
CursorBench 4.038.1%estimated ± 3.3 pp, high confidence
DeepSWE57.8%measured
FrontierCode 2.022.4%estimated ± 6.7 pp, medium confidence
HLE w/o tools38.2%estimated ± 4.3 pp, high confidence
PostTrainBench v1.127.7%estimated ± 4.4 pp, high confidence
SWE Multilingual80.5%measured
VulcanBench Frontier v478.1%estimated ± 5.0 pp, high confidence
ARC-AGI-30.0%estimated ± 8.3 pp, low confidence
CharXiv91.4%measured
CharXiv w/o tools83.5%measured
AI2D_TEST93.5%estimated ± 0.9 pp, low confidence
C-Eval94.6%estimated ± 0.9 pp, low confidence
CountBench97.3%estimated ± 1.3 pp, low confidence
DynaMath88.2%estimated ± 0.8 pp, low confidence
ERQA71.0%measured
INCLUDE86.6%estimated ± 2.0 pp, medium confidence
MathVision91.8%measured
MLVU (M-Avg)88.1%estimated ± 1.5 pp, high confidence
MMLU-ProX86.9%estimated ± 1.3 pp, medium confidence
MMLU-Redux94.9%estimated ± 1.7 pp, high confidence
MMMU-Pro80.1%estimated ± 3.0 pp, high confidence
NL2Repo48.9%measured
OmniDocBench 1.591.3%estimated ± 1.8 pp, medium confidence
PostTrainBench v128.7%estimated ± 4.6 pp, high confidence
RealWorldQA87.7%measured
ScreenSpot Pro82.6%estimated ± 4.8 pp, medium confidence
SimpleVQA75.2%estimated ± 7.3 pp, medium confidence
SuperGPQA76.6%estimated ± 7.2 pp, medium confidence
V*94.2%estimated ± 10.1 pp, low confidence
Video-MME (w/o subtitle)88.4%estimated ± 4.9 pp, low confidence
Video-MME (with subtitle)89.5%estimated ± 1.9 pp, high confidence
VideoMMMU87.6%estimated ± 1.5 pp, medium confidence
We-Math87.5%estimated ± 2.1 pp, low confidence
HealthBench Hard11.0%estimated ± 5.1 pp, low confidence
LiveCodeBench Pro86.2%estimated ± 6.0 pp, low confidence
MedXpertQA (MM)78.2%estimated ± 4.2 pp, medium confidence
MedXpertQA (Text)56.8%estimated ± 4.5 pp, medium confidence
MMLU-Pro (Arcee)87.2%estimated ± 3.4 pp, low confidence
PostTrainBench25.6%estimated ± 3.7 pp, high confidence
FrontierCode 1.1 Extended57.1%estimated ± 3.0 pp, medium confidence
HealthBench Professional53.1%estimated ± 4.2 pp, low confidence
ScarfBench62.6%estimated ± 3.8 pp, medium confidence
SimpleQA38.6%estimated ± 13.4 pp, low confidence
BabyVision w/ Python89.0%estimated ± 3.7 pp, low confidence
OpenHarmony Bench v1.055.8%estimated ± 2.7 pp, high confidence
BabyVision67.5%estimated ± 3.7 pp, high confidence
MMVU80.3%estimated ± 3.3 pp, medium confidence
PerceptionBench59.4%estimated ± 3.6 pp, medium confidence
ZeroBench16.2%estimated ± 3.6 pp, medium confidence
LVBench76.9%measured
RefCOCO (avg)91.2%estimated ± 3.1 pp, low confidence
OCRBench V274.0%estimated ± 5.4 pp, low confidence
BioBasicsBench24.7%estimated ± 2.2 pp, low confidence
MathVision w/ Python96.2%measured
SWE-Atlas Codebase QnA54.7%estimated ± 4.8 pp, high confidence